Hydrology – From Measurement To Hydrological Information (2)

CHAPTER 2
METHODS OF OBSERVATION

2.1 INTRODUCTION – THE HYDROLOGICAL CYCLE AS THE SUBJECT OF OBSERVATION

Water is found on Earth in signifi cant amounts in all three of its physical phases: liquid, solid and gaseous. It is also found in all three of Earth’s major environments that are readily accessible to humans:
the atmosphere, the seas and oceans, and the land masses. As water can readily move from one environment to another and can change from one phase to another in response to its environment, it is a
dynamic medium in both space and time.

The Earth’s system of repositories for the storage of water and the multitude of paths among the many repositories has been conceptualized as a cycle, as shown in Figure I.2.1. The science of hydrology has not traditionally encompassed the entire hydrological cycle, but has been limited to the land portion of the cycle and its interactions with the oceans and atmosphere.

Because humankind spends a predominant amount of time on the land surface and water is both a necessity for life and a potential hazard to it, hydrological knowledge is valuable in providing for our continuity and well-being.

One traditional means by which hydrological knowledge has been accumulated is through the measurement of the storage and flow of water at distinct
points in time and space.

Such measurements, also known as data, are analysed and synthesized to generate hydrological knowledge or information.
Volume II of this Guide deals with hydrological
analysis.

Two of the basic equations that describe the physics of the hydrological cycle are also pertinent in describing the systems that are used to make measurements of its transient properties: (a) the equation of continuity of mass; and (b) the equation of continuity of energy.

For example, one form of the equation of continuity of mass:

Q = AV

often serves as the basis for the determination of the flow rate in a stream or canal. In this equation, Q is the instantaneous rate of flow through a crosssection of the channel with the area, A and average flow velocity, V.

Figure I.2.1. The hydrological cycle

Traditionally, flow rate, also known as discharge, could not be measured directly for streams of even modest size. A cross-sectional
area, however, can be measured by sampling its spatial dimensions, and velocities can be sensed by the use of current meters. Thus, the use of this equation, described in detail in Chapter 5, has permitted
determinations of the rate of discharge of even the
largest rivers of the world.

Another example of the role of the equation of continuity of mass can be found in the observation of evaporation of water from a lake surface. In this example, the equation takes the form:


P + I – O – E = ΔS (2.2)

where P is the amount of precipitation that falls onto the surface of the lake during a period of observation, I is the inflow of surface water and groundwater during the period, O is the outflow of surface water and groundwater, E is the quantity of water evaporated from the surface of the lake during the period, and ΔS is the change in the volume of water in the lake during the period.


Precipitation can be measured according to the techniques described in Chapter 3; inflows and outflows can be measured using techniques
described in Chapters 4, 5 and 6; changes in the lake contents can be computed by relating the lakesurface elevations at the beginning and the end of the observation period to the respective contents at
those times; and water-level measurement is discussed in Chapter 5.

By measuring or otherwise observing four of the five terms in equation 2.2, the fifth, evaporation, can be computed algebraically.
Systematic hydrological observations are at the very core of the development of databases, information and knowledge required for the effective management of water resources.

This chapter discusses a number of subjects that are fundamental to the operation of hydrological and meteorological observing networks and to the production hydrological information.

The chapter provides an overview of hydrological standards and codes, accuracy of measurement concepts of network planning, methods of observation, measurement of physiographic characteristics, the role of hydrological data in information systems
and the linkages to sustainable development.

Some of these subjects are discussed in greater detail later
in this Volume. Where this is the case, cross-references are provided.


2.2 WATER RESOURCES INFORMATION
SYSTEMS

2.2.1 Need for data and information
The report of the International Conference on Water and the Environment (ICWE), held in Dublin in January 1992 (United Nations, 1992a), provides a compelling assessment of the importance of water
resources to the world’s environment and to its economy. Its specific policy statements highlight very effectively the role that hydrological services should play in achieving goals related to sustainable development.

ICWE addressed the following issues:
(a) Integrated water resources development and
management;
(b) Water resources assessment and impacts of
climate change on water resources;
(c) Protection of water resources, water quality and
aquatic ecosystems;
(d) Water and sustainable urban development and
drinking water supply and sanitation in the
urban context;
(e) Water for sustainable food production and rural
development, and drinking water supply and
sanitation in the rural context;
(f) Mechanisms for implementation and coordination at the global, national, regional and local
levels.


Volume II, Chapter 3, reviews the evolution of integrated water resources management and provides examples of best practices. The nature of the information that will be required to meet the needs of integrated water resources management is difficult to project. Perhaps, the best ideas can be gathered from considering recent trends in water management (2.2.4). Since data are gathered
for the use of water managers, whether in government or private agencies, changes in the way water is managed will influence the data and information demands.

The impacts of these changes may include: (a) Growing competition for water, resulting in a higher value being placed on available supplies and, ultimately, goods and services being redefined in terms of their water content – this could be exacerbated by declining water availability and quality in many areas;

(b) Economic pressures resulting in more user
fees, cost-sharing and local financing of water programmes, with a concurrent shift in emphasis from water development activities
to environmental programmes and demand management;
(c) Increased focus on water conservation and reuse in all phases of project development – in some areas, reclaimed water now costs less than freshwater supply;
(d) Environmental legislation designed to hold polluters and users accountable for their impacts on available supplies;
(e) Legal measures to ensure that users and water managers justify their needs, uses and management practices, and that increases priority be accorded to environmental water uses (for example, fish and wildlife habitat) versus the traditional economic uses (for example, agriculture and industry);
(f) Emphasis on basin and regional water planning
as a means of resolving transboundary issues
and disputes.

These trends indicate that greater coordination of data-collection efforts will be required to meet the needs of water managers in the future. Water management is becoming more integrated across
disciplines and specialities; therefore, compatible data on quantity and quality of surface water and groundwater, and on specific basins and regions will be required.

Current problems related to data accessibility, compatibility and
reliability will have to be resolved to meet these needs. In addition, water management challenges are closely linked to those of environmental management or ecosystem management.


Therefore, an increasingly holistic management approach is required. While many users will continue to need data for design and analysis purposes, increased attention must be paid to the need for comprehensive regional surface-water information that can be
applied to many different kinds of water issues and problems. This means that overview information, fact sheets and summaries, surface-water and precipitation mapping, hydrological assessments
of basins and regions, and water information relevant to the assessment of water quality and groundwater problems must be available. The use of real-time water data will continue to grow to
serve many needs.

2.2.2 Hydrological information systems

This Volume of the Guide deals with the field activities of operational hydrology. However, the data that are generated by the field activities are of little or no value if they cannot be readily and confidently accessed by the potential data users.

Operational hydrology within a given Hydrological Service can be considered as an information system providing a conceptual basis for the development of proper approaches, which ensure that the right
data are available in the right form at the right place and time.

Figure I.2.2 depicts the elements of a hydrological information system.

Ideally, the information system is embedded in a natural sequence of actions and decisions that begins with the perception of an opportunity and culminates in the implementation of decisions that maximize the net positive impacts of the opportunity.

A hydrological information system combined with a suite of numerical models – physical, statistical or socio-economic – comprises a decision support system.

With the decision support requirements firmly in mind, the designer of the information system can specify the procedures to be used to
analyse the hydrological data. These data-analysis technologies may be any one model or a combination of models that account for the probabilistic, stochastic or deterministic natures of the hydrological phenomena of interest. Volume II of this Guide (in particular Chapters 5 to 7) discusses many of these data-analysis technologies.


The actual data collection can begin at this point in the sequence, and it is also at this point that feedback, represented as dashed arrows in Figure I.2.2, begins to take place. All of the previous steps have been based on a specific level of knowledge about the hydrological conditions of interest. As data are collected, this level increases, and new data-analysis techniques and new network designs may
become appropriate. Guidance on data collection is given in 2.5.

Figure I.2.2. Components of a hydrological information system


From Figure I.2.2, it is possible to see that quality assurance is an integral phase of the information system that is relevant throughout the continuum from field activities to the dissemination of data
and information. Owing to its pervasive nature, quality-assurance guidance can be found throughout this Volume. No discussion of information systems is complete without mention of data management systems.

The information contained in a robust data-management system is available, not only for the uses for which the data were collected originally, but also for a multitude of uses that may never have been
anticipated. However, with robustness comes a price.

The options inherent in robust systems tend to make them difficult to use, as more training is required. This represents the first portion of the price.

This part of the cost can be minimized by user-friendly systems designs. The second cost factor is the potential loss of information that robustness entails. As a data-management system cannot be all things to all people, compromises must be made, which usually result in data compaction and loss of data attributes.

To reduce such loss, subsystems that retain more objective specific data can be appended to the robust, central system. Such systems are discussed in Chapter 10.

Current technology also allows the development of distributed hydrological information systems having searchable metadata. Provided that computer security matters are fully considered, such virtual data systems provide an effective and robust means
of accessing data and information required for decision-making.

The ultimate product of the information system is obtained by processing the data through the same data-analysis technology that was initially crucial in the design of the data network. The sequence
culminates by integrating the hydrological information into the decision process for which it was designed to have an optimal impact. The key to obtaining this optimality is the compatibility
among the decision technology, the data-analysis technology and the data network.

A well-designed information network contains synergism that is derived in three ways. First, information is a commodity that is not destroyed by its use. Thus, if properly preserved, it can be made available at minimal cost for many uses not anticipated at the time of its collection. Secondly, information can be used to improve understanding of hydrological processes. By improving process understanding, both the information content of the existing data
and all future data are increased.

Thirdly, synergism evolves by taking advantage of the accomplishments of others. New approaches and technologies for the design of information systems, like the data they
contain, are recyclable commodities.


2.2.3 Uses of water resources information

Hydrological or Hydrometeorological Services or related agencies have been established in countries for systematic water resources data collection, archiving and dissemination described elsewhere in
this Volume. Their primary role is to provide information to decision-makers on the status of and trends in water resources.

Such information may be required for the following purposes (WMO/
UNESCO, 1991):
(a) Assessing a country’s water resources (quantity,
quality, distribution in time and space), the
potential for water-related development and
the ability of the supply to meet actual or foreseeable demand;
(b) Planning, designing and operating water
projects;
(c) Assessing the environmental, economic and
social impacts of existing and proposed water
resources management practices and planning
sound management strategies;
(d) Providing security for people and property
against water-related hazards, particularly
floods and droughts;
(e) Allocating water among competing uses, both
within the country and cross-border;
(f) Meeting regulatory requirements.
Most frequently, water resources information has
been collected for a specific purpose, such as the
design of a hydroelectricity scheme.

However, increasing competition among users for scarce
water requires that resources be managed in an
integrated fashion, so that the interactions among
several projects and users may be understood.

This places a much greater burden on the suppliers of
water resources information, because a variety of
types of information is simultaneously needed, and
has to be presented in different forms for different
users.

This makes it essential that assessment agencies understand the needs of all their users, and not just those with whom they have traditionally dealt.


Even more demanding is the need to look ahead to
the possible needs of future users of data and to
commence collecting the information before an
actual demand can be demonstrated with certainty.

Therefore, it is necessary that the design and updating of data-collection networks, especially the principal stations, be coordinated to ensure that stations for monitoring the various elements of the
water cycle are suffi ciently related, both in number and location, to achieve an integrated network


(2.4). Such an approach would enhance the information content of the data sets for both current and unforeseen future needs.


With the growing recognition of such issues as
global climate change and the environmental impacts of human activities, such as urbanization, there is an increasing emphasis upon the information required as a foundation for sustainable development and management of water resources.

Volume II, Chapter 3, describes the rationale for integrated water resources management and presents elements of best practice.


2.2.4 Types of water resources
information

The diversity of possible uses of water resources information implies that there is a considerable range of types of data. Conventional water resources information comprises the statistics of a variety of
meteorological and hydrological elements.

The elements include the following (WMO/UNESCO,
1991):
(a) Precipitation, for example, rainfall, snow and
fog drip;
(b) River levels and fl ows, and lake and reservoir
levels;
(c) Groundwater levels;
(d) Evapotranspiration;
(e) Sediment concentrations and loads in rivers;
(f) Water quality (bacteriological, chemical; and
physical) of surface water and groundwater.
The statistics include the following:
(a) Mean annual, monthly, or seasonal values;
(b) Maxima, minima and selected percentiles;
(c) Measures of variability, such as the standard
deviation;
(d) Continuous records in the form, for example,
of a river flow hydrograph.

There is a requirement for both historical and realtime data to cater to the full range of needs from water resources planning to project design and flood warning.

Flood or low flow forecasting (Volume II, Chapter 7) may require data to be synthesized for the future by using numerical fl owrouting models (Volume II, 6.3.4).


The Water-resource Assessment Activities: Handbook for
National Evaluation (UNESCO/WMO, 1988) recognizes a number of types of water resources projects for which hydrological information is required, as given in Table I.2.1.


Together, these imply a vast range of water-related data and information that the Hydrological Services and other related agencies may be required to collect and archive. Different countries have different priorities that depend on their level of economic and social development, the sensitivity of their natural environments to
disturbance by human activity, and the nature of the physical environment itself, for example, climate, topography and the abundance or otherwise of water.

There are several critical requirements for an effective water resources assessment programme:
(a) Data of high quality must be collected to permit
confident statistical analysis;
(b) The data and the information that they provide
must be carefully targeted to the requirements
of users;
(c) An integrated observation programme, in
which measurements of several variables are
made simultaneously, is required to provide the
greatest total value;
(d) Other forms of information that are compatible
with and can be analysed with water resources
information should be available;(e) An effective system is needed for archiving
and disseminating data to ensure that they
are not lost or corrupted and that they are
made avail able in a form that enables analysis
(Chapter 10).


The above requirements can be met by the application of contemporary technologies – for example,
telemetry, to make data available in near-real time –
by implementing searchable computer databases,
by remote-sensing to collect areal information more
effectively and by Geographical Information
Systems (GIS) (2.6.7) to provide a means of analysing spatial data. At the same time, new computer
storage devices and the use of the Internet make the
data more readily available. Nevertheless, technology is not the only requirement, and a trained and
well-managed staff is of even more fundamental
importance.

As financial resources become increasingly limited in many countries, it becomes ever more vital that effective organizational structures
are in place to ensure that those resources are used
most efficiently.


In addition to the more conventional measurements,
there is a growing recognition of the need to measure other aspects of the freshwater environment and
of the wider environment in which freshwater is
only a single component. These include:


(a) The volumes of water needed for industrial,
domestic and agricultural use, and for navigation. These are now signifi cant modifi ers of the
hydrological cycle in many basins;
(b) Attributes of rivers and required volumes of
water related to instream uses, for example,
freshwater fi shery habitats or recreation;
(c) Watershed characteristics that may be related
to hydrology, for example, vegetation patterns,
soil moisture, topography and aquifer
characteristics;
(d) Environmental concerns, for example,
eutrophication of lakes and damage to natural
freshwater and estuarine ecosystems.

Table I.2.1. Hydrological information required for water resources projects

HYDROLOGICAL SYMBOLS, CODES
AND ACCURACY OF MEASUREMENTS
2.3.1 Units and symbols

Standardization of units and symbols is desirable
and can be achieved through the use of those
recommended in Tables I.2.2–I.2.4 (ISO, 1993).
Commonly used units and conversion factors are
also given. All symbols and units used in the Guide
conform to those in the tables.


2.3.2 Hydrological codes
2.3.2.1 General

All systems for the transmission of data make use of
some form of coding method, the purpose of which
is to ensure that the information is transmitted
quickly and reliably (9.3). In the case of fully
automated systems, the information must be put into
coded form before being processed. For these reasons,
the codes are composed of standard forms that enable
the information to be transmitted and given in a
form compatible with processing. Such processing is
usually preceded by quality control (9.8).


The structure of international codes is governed by
convention, and codes are developed as a collective
effort. For a long time, WMO has been developing
codes to meet requirements for the exchange of
meteorological data.


In operational hydrology, requirements for data are
not on a worldwide scale and yet there have been a
multiplicity of codes introduced in this fi eld. This
led the WMO Commission for Hydrology to develop
international hydrological codes. The purpose of
these codes is to cover general requirements so that,
as far as possible, the procedures for coding and
collecting hydrological data are standardized.

The HYDRA and HYFOR codes, which were developed
and used in the past, are no longer recommended
for use. Instead, the character form for the representation and exchange of data (CREX) code has been
developed in recent years for use in the representation and transmission of hydrometeorological data.


This code may be found to be particularly useful in
the case of large national and international basins,
in which a large number of stations are connected
to a processing centre. The observations are coded,
usually manually, by an observer, and then transmitted to a collecting centre for processing.


2.3.2.2 Character form for the
representation and exchange of data
CREX is the name of a character code for the representation and exchange of meteorological, hydrological and water-quality data.

Table I.2.2. Recommended symbols, units and conversion factors
  • Note: Where international symbols exist, these have been used where appropriate and are indicated as ISO in the last column.
  • Column IV = Conversion factor (Column VI) x Column V
    ** General terms. For detailed terminology and symbols, see the Guide to Meteorological Instruments and Methods of Observation (WMO-No. 8).
Table I.2.3. Miscellaneous symbols
Table I.2.3. Miscellaneous symbols
Table I.2.4. Recommended units as appearing in Table I.2.2.

Although originally designed for the exchange of data for which there were no suitable existing WMO code forms, CREX has been used recently as standard code form for data transmitted from data-collection platforms (DCPs).

A CREX message shall consist of one or
more subsets of related meteorological data defined, described and represented by a single CREX entity. For observational data, each subset shall correspond to one report. CREX uses many of the principles of the previous BUFR code, and each message consists
of sections as follows:

Further information can be found at https://www.
wmo.int/pages/prog/www/WMOCodes.html
.

2.3.3 Accuracy of hydrological
measurements
2.3.3.1 Basic principles


Theoretically, the true values of hydrological elements
cannot be determined by measurements because
errors of measurement cannot be eliminated
completely. The uncertainty in measurement has a
probabilistic character that can be defined as the
interval in which the true value is expected to lie with
a certain probability or confidence level.

The width of the confidence interval is also called error band.
If measurements are independent one from the other, the uncertainty in the results of measurements can be estimated by taking at least

20–25 observations and calculating the resulting
standard deviation, and then determining the
confidence level of the results. This procedure
cannot usually be followed in hydrometric
measurements, because of the change in the value
to be measured during the measuring period. For
instance, many consecutive measurements of
discharge with current meters at constant stage is
clearly impracticable in fi eld conditions. Thus an
estimate of the uncertainty has to be made by
examining the various sources of errors in the
measurement.


Another problem in applying statistics to hydrological data arises from the assumption that observations are independent random variables
from a fi xed statistical distribution. This condition
is seldom met in hydrological measurements. River
fl ow is, by nature, not purely random. It depends
on previous values. It is generally accepted that
some aspects of the departure of hydrological data
from the theoretical concept of errors is not serious. However, it should be stressed that no
statistical analysis can replace correct observations, in particular because spurious and systematic
errors cannot be eliminated by such analysis. Only
random errors can be characterized by statistical
means.


Section 2.3.3 contains definitions of basic terms
related to the accuracy of hydrological measurements. Methods for estimating uncertainty are
introduced and numerical values of accuracy, required
for the most important hydrological parameters, are
given. References to the existing recommendations
contained in the Technical Regulations (WMO-No. 49)
and other publications are also included.


2.3.3.2 Definitions of terms related to
accuracy


The definitions of the terms related to accuracy given
below take into account those given in the Technical
Regulations (WMO-No. 49), Volume III – Hydrology
and in the Guide to Meteorological Instruments
and Methods of Observation (WMO-No. 8):
Accuracy: The extent to which a measurement agrees
with the true value. This assumes that all known
corrections have been applied.


Confidence interval: The interval which includes the
true value with a prescribed probability and is estimated as a function of the statistics of the sample
(Figures I.2.3 and I.2.4).
Confidence level: The probability that the confidence
interval includes the true value (Figures I.2.3 and
I.2.4).
Correction: The value to be added to the result of a
measurement to allow for any known systematic
error and thus obtain a closer approximation to the
true value.

Figure I.2.3. Explanation of errors
Figure I.2.4. Explanation of errors in linear regression

Error: The difference between the result of a measurement and the true value of the quantity
measured. This term is also used for the difference
between the result of a measurement and the best
approximation of the true value, rather than the
true value itself. The best approximation may be a
mean of several or many measurements.
Expected value: The best approximation of the true
value, which may be a mean of several, or of many
measurements.


Hysteresis (instrument): That property of an instrument whereby it gives different measures, for the
same actual value, according to whether that value
has been reached by a continuously increasing or
continuously decreasing change of the variable.


Measurement: An action intended to assign a number
as the value of a physical quantity in stated units.
The result of a measurement is complete if it
includes an estimate (necessarily in statistical terms)
of the probable magnitude of the uncertainty.
Normal distribution: A mathematically
defined, symmetrical, bell-shaped, continuous
distribution, traditionally assumed to represent
random errors.


Precision: The closeness of agreement between
independent measurements of a single quantity
obtained by applying a stated measurement procedure several times under prescribed conditions.


Accuracy has to do with closeness to the truth,
precision has to do only with closeness together.


Precision of observation or of reading is the smallest unit of division on a scale of measurement to
which a reading is possible either directly or by
estimation.


Random error: That part of the error that varies in an
unpredictable manner, in magnitude and in sign,
when measurements of the same variable are made
under the same conditions (Figure I.2.3).
Range: The interval between the minimum and
maximum values of the quantity to be measured, for which the instrument has been constructed, adjusted or set. It can be expressed
as a ratio of maximum and minimum measurable values.
Reference measurement: A measurement utilizing the
most advanced state of science and the latest technologies. The result of the reference measurement
is used to estimate a best approximation to the true
value.


Repeatability: The closeness of agreement, when
random errors are present, between measurements
of the same value of a quantity obtained under the
same conditions, that is, the same observer, the
same instrument, the same location, and after intervals of time short enough for real differences to be
insignificant.


Reproducibility: The closeness of agreement between
measurements of the same value of a quantity
obtained under different conditions, for example,
different observers, instruments, locations, and

after intervals of time long enough for erroneous
differences to be insignificant.


Resolution: The smallest change in a physical variable that causes a variation in the response of a measuring system.


Sensitivity: The relationship of the change of the
response to the corresponding change of the stimulus, or the value of the stimulus required to produce a response exceeding, by a specified amount, the response already present due to other causes.

Spurious value: Value known for certain to be in
error, for example, due to human mistakes or instrument malfunction

(Figure I.2.3). Standard deviation (Sy): This is a measure of the
dispersion of values about their mean. It is defined
as the positive square root of the sum of the squares
of the deviation from the arithmetic mean, divided
by (n – 1). It is given by:

where y
_
is the arithmetic mean of the sample of n
independent measurement of the variable y, and
(n – 1) indicates the loss of one degree of freedom.
Standard error of estimate (Se ): A measure of the variation or scatter of the observations about a linear
regression. It is numerically similar to the standard
deviation except that the linear-regression relation
replaces the arithmetic mean and (n – 1) is replaced
by (n – m):

where d is the deviation of an observation from the
computed regression value, m is the number of
constants in the regression equation, and (n – m)
represent the degrees of freedom in the equation
derivation.


Systematic error: That part of the error that
either:
(a) Remains constant in the course of a number
of measurements of the same value of a given
quantity; or
(b) Varies according to a definite law when conditions change (Figure I.2.3).

Tolerance: The permissible accuracy in the measurement of a specified variable.

Tolerance limit: The limiting lower or upper value
specified for a quantitative characteristic.
True value: The value that characterizes a quantity
in the conditions that exist at the moment when
that quantity is observed. It is an ideal value that
could be known only if all causes of error were
eliminated.


Uncertainty: The interval about the measurement
within which the true value of a quantity can be
expected to lie with a stated probability
(Figure I.2.3). The numerical value of uncertainty
is a product of the true standard deviation of the
errors and a numerical parameter depending on
the confi dence level:
e = ± ασ
y ≈ αsy (2.5)


The standard deviation, sy, computed from n
observations, approaches the true standard deviation, σy, as n approaches infi nity. In the case of normal distribution of error, numerical parameters are as follows:

2.3.3.3 Types of error


Spurious errors should be eliminated by discarding
the values of measurements concerned.

These errors can be identifi ed by a statistical-outlier
test, such as the one described in ISO 5168 (ISO,
2005) that gives a rejection criterion.

Systematic error originates mainly from instrumentation and cannot be reduced by increasing the number of measurements if the instruments and conditions remain unchanged. If the systematic error has a known value, this value should be added
to or subtracted from the result of the measurement, and error due to this source should be considered zero. Systematic error should be eliminated by correcting, properly adjusting or changing
the instrument, and/or by changing the flow conditions, for example, the length of straight approach channel of a stream-gauging section.

These errors are often due to difficult measuring conditions,
such as unsteady fl ow, meandering and bad location of stations.
Random errors cannot be eliminated, but their effects
can be reduced by repeated measurements of the
element. The uncertainty of the arithmetic mean
computed from n independent measurements is
several times smaller than the uncertainty of a single
measurement. The distribution of random errors can
usually be assumed to be normal (Gaussian). For
certain cases, normal distribution can or should be
replaced by other statistical distributions.


2.3.3.4 Sources of error
Each instrument and measuring method has its
own sources of error. Therefore, it would be difficult
to list all possible sources of error. The specific
sources are usually mentioned in the descriptions
of the design of the instruments and operating
procedures, such as those in ISO Standards, and the
Manual on Stream Gauging (WMO-No. 519). Some
typical sources of error include:


(a) Datum or zero error originates from the incorrect determination of the reference point of an
instrument, for example, staff-gauge zero level,
difference between the staff-gauge zero and the
weir-crest levels;
(b) Reading error results from the incorrect reading
of the indication by the measuring instrument,
for example, due to bad visibility, waves, or ice
at the staff gauge;
(c) Interpolation error is due to inexact evaluation
of the position of the index with reference to
the two adjoining scale marks between which
the index is located;
(d) Observation error is similar to the reading error
and is attributed to neglect or incompetence of
the observer;
(e) Error due to the negligence of one or several
variables needed to determine the measured value (for example, assuming a unique
stage-discharge relationship during periods of
unsteady fl ow when slope as well as stage is a
signifi cant determinant of discharge);
(f) Hysteresis (defi nition under 2.3.3.2);
(g) Non-linearity error is that part of error whereby
a change of indication or response departs from
proportionality to the corresponding change
of the value of the measured quantity over a
defi ned range;
(h) Insensitivity error arises when the instrument
cannot sense the given change in the measured
element;
(i) Drift error is due to the property of the instrument in which its measurement properties
change with time under defi ned conditions of
use, for example, mechanical clockworks drift
with time or temperature;
(j) Instability error results from the inability of an
instrument to maintain certain specifi ed metrological properties constant;
(k) Out-of-range error is due to the use of an instrument beyond its effective measuring range,
lower than the minimum or higher than the
maximum value of the quantity, for which the
instrument/installation has been constructed,
adjusted, or set (for example, unexpected high
water level);
(l) Out-of-accuracy class error is due to the
improper use of an instrument when the minimum error is more than the tolerance for the
measurement.


2.3.3.5 Secondary errors of measurement
Hydrological observations are often computed
from several measured components. For example,
discharge at measuring structures is computed as
a function of a discharge coeffi cient, characteristic dimensions and head. For estimating the
resultant uncertainty, the Gauss error transfer
(propagation) theory can be applied.

Resultant uncertainty is often referred to as overall
uncertainty, which can be calculated from the
uncertainties of the individual components if the
errors of the individual components are assumed to
be statistically independent.


If a quantity, Q, is a function of several measured
quantities, x, y and z, the resultant uncertainty,
eQ, in Q due to uncertainties, ex, ey and ez, in x, y
and z, respectively, should be evaluated by the
simplified equation of the transfer
(propagation):

where ∂Q/∂x, ∂Q/∂y and ∂Q/∂z are the partial differentials of the function expressing explicitly the
relationship of the dependent variable with the
independent variables.

In hydrological measurements, it is very rare that a
measurement can be repeated under the same
conditions in the field. The standard deviation
should therefore be determined by using data of
changing variables as in the case of a discharge
rating curve. The standard error of estimate:

of the mean of observations is extremely important for characterizing the stage-discharge
relationship, which needs special treatment
because the relationship is not linear, but approximately logarithmic. It is an estimate of the
accuracy of the computed mean relationship in a
regression and, therefore, it is the range within
which the true mean would be expected to lie
(Figure I.2.4).


For small samples, it could be useful to have a
corrected standard error of estimate, obtained by
multiplying seen by n 1/2 , resulting as:

2.3.3.6 Characterization of instruments and
methods of observation


The accuracy of a measuring instrument can be
characterized by an uncertainty at a given value,
corresponding to a maximum or minimum measurable value. The accuracy of an instrument without a reference value can be misunderstood or misinterpreted. The instrument accuracy is in many cases only one component of the overall accuracy of the
measurement.


For characterization of uncertainty, the 95 per cent
confidence level is commonly used. That is, in 5 per
cent of the cases, the error could be outside the
stated confidence interval. According to the
Technical Regulations (WMO-No. 49), Volume III,
measurement uncertainties should be reported in
one of the following forms:
(a) Uncertainties expressed in absolute terms:
Measured value of hydrological elements, for
example, discharge: Q = …
Random uncertainty: (er
)95 = …
(b) Uncertainties expressed in percentage terms:
Measured value of the hydrological elements
Q = …
Random percentage uncertainty (er
)95%= …%
In practice, uncertainties of measurements are given
in a form where uncertainty is expressed as a ratio
(or percentage) of Qm, the measured value. For
example, in the case of (er
)95 = 10%, Qm ± 0.10 Qm
will contain the true value of Q 95 per cent of the
time. In this case, the uncertainty is formulated by
assuming average measurement conditions.
2.3.3.7 Recommended accuracy of
hydrological measurements
The recommended accuracy depends mainly on the
anticipated use of the measured data (the purpose
of the measurement), on the potentially available
instruments and on the available fi nancial resources.
Therefore, it cannot be a constant value. Rather it
should be a fl exible range. The recommended accuracy levels are tabulated in Table I.2.5 as a general
guidance for instruments and methods of observation. In many countries, national standards regulate
the required accuracies.


2.3.4 Calibration of instruments
One of the major sources of error, as stated above, is
due to change in measurement characteristics of
the instruments. Hydrological instrumentation
comprises a large variety of mechanical, electromechanical and electronic devices. Mechanical
instruments such as current meters or anemometers
provided by reputable manufacturers are made with
precision dies and are usually supplied with a
factory calibration table. The factory calibration
will, of course, only apply if the instrument is not
damaged in use and is properly maintained. Many
national hydrological agencies operate facilities to
verify factory calibrations and international standards for the manufacture and calibration of, for
example, current meters.


Increasingly, there is a trend towards replacing
mechanical devices with electronic ones. Although
they are more reliable than mechanical devices,
they usually are not repairable in the field and must
simply be substituted for a replacement device.


Electronic instrumentation poses particular problems for hydrological agencies that may be making a transition from electromechanical devices to electronic ones as the calibration issues may be quite
different. Calibration of an electronic instrument may drift due to temperature or pressure changes, or solid-state sensors may become fouled during use. It is essential that instruments be designed to
function in the range of conditions that are likely to occur at the data-collection site. Some instruments have built-in calibration checks and it is important that these be used.

Table I.2.5. Recommended accuracy (uncertainty levels) expressed
at the 95 per cent confi dence interval

Notes:

  1. When a range of accuracy levels is recommended, the lower value is applicable to measurements under relatively good conditions and the higher value is applicable to measurements under difficult situations.
  2. Obtaining the recommended accuracy of precipitation measurements, 3–7 per cent, will depend on many factors, including gauge
    characteristics. For gauges having their orifice above the ground, the gauge catch deficiency is strongly determined by wind speed and
    precipitation type. The catch deficiency for light snow falling during strong wind can for example be 50 per cent or more.

2.4 DESIGN AND EVALUATION OF
HYDROLOGICAL NETWORKS


2.4.1 General concepts of network design


A hydrological data network is a group of data collection activities that is designed and operated to address a single objective or a set of compatible objectives. Frequently, the objectives are associated
with a particular use that is anticipated for the data
being collected in the network – for example, for a
water resources assessment, a development plan, or
a project design. A particular hydrological station
or gauge may be included in more than one network
if its data are being used for more than one purpose.
In most parts of the world this is more commonly
the case than not.

Alternatively, a single network may consist of several types of station or gauge if they are all contributing information to the
network’s objective. For example, both raingauges
and stream gauges might be included in a flood
forecasting network.


The term network is frequently used in a less rigorous sense. It is often possible to hear of surface-water network, groundwater network, precipitation network, or water-quality network when the speaker is referring to an aggregation of gauges and stations
that have no coherence in their objectives.

Data collection sites included in a network under this looser definition may even have disparate uses for the data being collected. This disparity of usage is more than just a semantical oddity. It can cause
confusion and false expectations when network analysis and design are being discussed among programme managers and hydrologists.


A network design could be based on a maximization of the economic worth of the data that are to be collected. However, such is not the case in the real world. Generally, in water resources decisionmaking, the economic impacts of hydrological data
are never considered. Decisions are made based on
the available data; the option of delaying the decision to collect more data is not explored, or deemed unacceptable. However, several examples of exceptions to this general rule are contained in the
Cost–benefit Assessment Techniques and User Requirements for Hydrological Data (WMO-No. 717) and in the Proceedings of the Technical Conference on the Economic and Social Benefits of Meteorological and Hydrological Services (WMO-No. 733). A review of the hydrometric network in one Canadian province
indicated that the cost–benefit ratio of the existing provincial network was 19 and that the network could be tripled in size to maximize economic benefi ts (Azar and others, 2003).

Even in nations with very dense hydrometric networks, such as the
United Kingdom, economic analysis inevitably
demonstrates that benefi ts of hydrometric networks
exceed the cost (CNS, 1991). Nonetheless, many
countries suffered considerable reductions in their
hydrological networks in the 1980s and 1990s as a
consequence of budget reductions for monitoring
agencies (Pearson, 1998).

For example, network reductions in Canada, Finland, New Zealand
and the United States of America were 21, 7, 20 and 6 per cent, respectively. Network reductions, with rare exceptions such as New Zealand, continue.


In lieu of complete economic analyses, network
designs are usually based on surrogate measures of
the economics or on guidance such as that presented
subsequently in this chapter.


2.4.1.1 Definition of network design


A complete network design answers the following
questions pertaining to the collection of hydrological data:
(a) What hydrological variables need to be
observed?
(b) Where do they need to be observed?
(c) How often do they need to be observed?
(d) What is the duration of the observation
programme?
(e) How accurate should the observations be?
To answer these questions, network design can be
conceptualized as a pyramid, as shown in
Figure I.2.5. The base of the pyramid is the science
of hydrology. Without a thorough understanding
of the hydrological setting of the area in which the
network is to be established, there is little chance
that the resulting network will generate information in an effective manner.

Hydrological understanding comes from both education and
experience, but there is no substitute for experience
when initiating a hydrological network in an area
where little or no historical data are available.
The right-hand side of the pyramid deals with
quantitative methods for coping with hydrological
uncertainty. Because of measurement errors and
errors caused by sampling in space and time, there
will always be hydrological uncertainty. Perfect
hydrological information can never exist.
Probabilistic descriptions of these errors are the
most effective means of dealing with the resulting
uncertainty. Probability theory provides the
theorems and the language for doing so and also
yields the understanding that is necessary for
appropriate use of the tools of statistics.

Figure I.2.5. The basic building blocks of
network design

In Figure I.2.5, statistical tools are represented by
sampling theory and by correlation and regression
analyses, which are commonly used in quantitative
network-design approaches.

However, there are many other branches of statistics that may be found useful in network analysis and design. The capstone
of uncertainty is Bayesian analysis, which pertains
to the level of uncertainty in the descriptions of hydrological uncertainty. In other words, the probabilistic descriptions of uncertainty, based on statistics of finite samples of hydrological
data, are uncertain in themselves. Reduction of
uncertainty about uncertainty is a key aspect of
taking full advantage of the information
contained in the data that the network will
generate.

The column in the middle of the structure, labelled
optimization theory, is often included taxonomically as a part of socio-economic analysis. However, even in the absence of socio-economics, the optimization theory is often used in hydrological
network design. Thus, it is included here as a separate component of the structure. A suite of
mathematical programmes, each with its own utility and shortcomings, comprises optimization
theory, which is often referred to as operations
research. The context of the network-design problem determines which, if any, of the mathematical
programmes can be used in a given situation. Often,
the choice between two or more network designs
must be made on the basis of judgement because
appropriate optimization tools either do not exist
or are too consuming of computer resources to be
efficient.

Atop the pyramid is decision theory, which is a
formal mechanism for integrating all of the underlying components. The application of decision
theory in network design is not required – it is not
even possible in most circumstances. However, an
understanding of its pretexts and premises can
make a network designer more cognizant of the
impacts of his or her final decisions.


The left-hand side of the pyramid represents a
rather amorphous group of technologies under the
heading of socio-economic analysis. In addition to
social sciences and economics, this part of the
network-design structure also encompasses policy
science and even politics. The latter plays a very
important role in the realization of the potential
benefi ts of water and, thus, also in the ultimate
value of the data from the network. The left-hand
side of the structure is the part that usually receives
little rigorous consideration in the design of the
data network. This is probably attributable to two
causes: the subject matter is difficult to treat in an
objective, mathematical way; and to do so in a
substantive manner requires the synthesis of
inputs from many disciplines beyond those of
hydrology and water resources engineering. Thus,
a network design that includes a significant socioeconomic analysis will probably be both expensive
and time-consuming.


Nevertheless, hydrological data-collection sites
are often installed to meet pressing social needs
and economic constraints with relatively little
thought to meeting long-term hydrological
information needs. Aside from meeting scientific needs, data-collection sites may be installed
to assist water mangers in responding to extreme
events such as floods or droughts, allocating
water supplies among competing uses, or meeting regulatory requirements. Sites operated for
these latter purposes may also lead to increased
hydrological understanding, but the resulting
network is by no means optimized for that
purpose.


2.4.1.2 Surrogate approaches
Since full-scale and complete network design is
either impossible or impractical in today’s world,
approaches that substitute surrogate measures,
objectives, or criteria are actually used to answer
the questions that comprise network design. For
example, a common substitution is to maximize
information content from a network in lieu of
optimizing the economic value of the data. Studies
have shown that, if information is used properly,
it can be expected to contribute to the economic
worth resulting from a decision. The more information, the better the decision. However, the
economic impact of information is not linearly
related to its magnitude and the marginal worth of
additional information decreases with the amount
of information that is available. Thus, the use of
this surrogate criterion can lead a Hydrological
Service in the right direction if only sparse hydrological information is available, but its use can
cause the collection of excess data if the region of
interest already has a reasonably adequate information base.

Among the basic analytical techniques that take
advantage of surrogates in the design of networks
are cartographic analysis, correlation and
regression methods, probabilistic modelling,
deterministic modelling and regionalization
techniques. Each method has particular
applications and the choice depends on the limitations of available data and the type of problem
under consideration.

Quite often the different techniques are combined in certain applications.

The Casebook on Hydrological Network Design
Practice (WMO-No. 324) presents applications of
these techniques as a means of determining
network requirements. Further examples are
contained in other publications (WMO/IHD
Project Report No. 12; WMO-Nos. 433, 580,
806)

2.4.1.3 The basic network
The worth of the data that derive from a network is
a function of the subsequent uses that are made of
them. Nevertheless, many of the uses of hydrological data are not apparent at the time of the network
design and, therefore, cannot be used to justify the
collection of specific data that ultimately may be of
great value. In fact, few hydrological data would be
collected if a priori economic justifi cations were
required. However, modern societies have developed a sense that information is a commodity that,
like insurance, should be purchased for protection
against an uncertain future. Such an investment in
the case of hydrological data is the basic network,
which is established to provide hydrological information for unanticipated future water resources
decisions. The basic network should provide a level
of hydrological information at any location within
its region of applicability that would preclude any
gross mistakes in water resources decision-making.
To accomplish this aim, at least three criteria must
be fulfi lled:
(a) A mechanism must be available to transfer
the hydrological information from the sites at
which the data are collected to any other site in
the area;
(b) A means for estimating the amount of hydrological information (or, conversely, uncertainty)
at any site must also exist;
(c) The suite of decisions must include the option
of collecting more data before the fi nal decision
is made.
2.4.1.3.1 The minimum network
In the early stages of development of a hydrological network, the first step should be the
establishment of a minimum network. Such a
network should be composed of the minimum
number of stations which the collective experience of hydrological agencies of many countries
has indicated to be necessary to initiate planning
for the economic development of the water
resources.


The minimum network is one that will avoid serious deficiencies in developing and managing
water resources on a scale commensurate with
the overall level of economic development of the
country. It should be developed as rapidly as
possible by incorporating existing stations as
appropriate. In other words, this pragmatic
network will provide the basic framework for
network expansion to meet future needs for
specifi c purposes. It is emphasized that a minimum network will not be adequate for the formulation of detailed development plans and
will not meet the numerous requirements of a
developed region for the operation of projects
and the management of water resources.


2.4.1.3.2 Expanding the information base
Once the minimum network is operational, regionalized hydrological relationships, interpreted
information and models can be formulated for estimating general hydrological characteristics,
including rainfall and runoff at any location in the
area. The basic network of observing stations should
be adjusted over time until regional hydrological
relationships can be developed for ungauged areas
that provide the appropriate level of information.
In most cases, this adjustment will result in increases
in the densities of hydrological stations. However,
this is not always the case. Since models are used to
transfer the information from the gauged to the
ungauged sites, the quality of the model is also a
factor in determining the density of the basic
network. If a model is particularly good, it can distil
the information from the existing data better than
a poorer model, and the better model would require
less data to attain a given level of regional information than would the poorer one. In an extreme
situation, the regional model might be so good that
the level of data collection in the basic network
could be reduced.


Owing to the broad dependence on the stations in
the basic network, it is very important that the
records from all of these stations be of high quality.
Even if the installation of a station is adequate, its
records may be of little value if it is not operated
correctly. Continuous operation may be difficult –
especially over a period of 20 years or more. A
minimum network, in which stations are abandoned or irregularly observed, will have its effective
density reduced and is, therefore, no longer an
adequate minimum network. For that reason, care
should be taken not only in establishing, but also in
providing for, the continuing operation of these
stations and for monitoring the reliability and accuracy of the collected records.


Economic, as well as technical considerations, are
involved in the design and implementation of basic
networks, and the number of stations requiring
observation over an indefi nitely long period cannot
be excessive. Consequently, a sampling procedure
may be adopted to maximize the cost-effectiveness
of the basic network. One such approach categorizes the stations as either principal or base stations,
or secondary stations. The secondary stations are
operated only long enough to establish a stable relationship, usually by means of correlations, with
one or more of the base stations. A new secondary
station can then be established with the equipment
and funds that had been in use at the discontinued
site. Records can be reconstructed at the discontinued site by means of the base-station records and
the inter-station relationship. At times, it may be
necessary to re-establish secondary stations if it is
believed that the conditions either at the secondary
site or at its related base station(s) have changed.
The perpetual nature of the principal stations in the
basic network provides a basis for monitoring longterm trends in hydrological conditions in the
region. This is particularly important in the light of
potential changes in the hydrological cycle that
could be caused by land-use changes or by increases
in stratospheric greenhouse gases.


2.4.1.4 Integrated network design
The hydrological cycle is a continuum, and its interconnections permit the partial transfer of
information obtained in one part of the cycle to
another. The efficiency of such transfers is proportional to the degree of hydrological understanding
that is captured in the models that are used to route
the water (and the information) between the parts
of the cycle. For example, precipitation records on
or near a gauged drainage basin permit the reconstruction of streamflow records during periods
when the stream-gauge malfunctions if a valid
precipitation-runoff model has been calibrated
during times when all gauges were functioning
properly. A groundwater observation well may
perform a similar role for malfunctions of the
stream gauge if the well is monitoring the water
table of an aquifer that is directly connected to the
stream.


To date, little has been done to include these interactions in network designs in an explicit manner. Ideally, the complementarity between the raingauges and the stream gauges that are operated in a
fl ood-forecasting network could be used in designing a network for water resources assessment, for
example. If the economic trade-offs between the
two networks could be defi ned, they could be optimized together and peak effi ciencies in information
generation could be attained for both. In spite of
this technological shortcoming, networks should
be designed iteratively, and the outcomes of an
existing design should become starting points for
subsequent designs. By extension of the above
example, this can be illustrated. The flood-forecasting network will probably have stream gauges and
precipitation gauges at rather specific locations to
meet its information needs. As the water resources
assessment will generally have less specific requirements for its information sources, it will be likely
that many of the gauges of the fl ood-forecasting
network can be incorporated into the assessment
network and used as initial given conditions for its
design. This iterative approach is particularly useful
when designing generalized networks, like the basic
network on the basis of networks, with more restrictive information demands. Networks with more
restrictive demands include benchmark stations,
representative basins and networks for operational
purposes.


2.4.1.4.1 Stations for operational purposes
Stations may be established for such specific
purposes as reservoir operation, irrigation, navigation, water-quality monitoring or fl ood forecasting.
Benchmark or reference stations would also belong
to this category. The length of operation of special
stations is related to the purpose for which they
were installed.


In some cases, the specifi c purpose to be served
may require observations on only one particular
aspect of an element, or be confined to one season
of the year. For example, a hydrometric station
may consist of a crest gauge for recording only the
maximum flood peak or a storage gauge for measuring the total precipitation during a season.
Although such stations may perform a valuable
function, they do not provide the data required
for general hydrological analyses. Consequently,
such stations may or may not be included in a
basic hydrological network.


2.4.1.4.2 Benchmark stations
Each country and each natural region of large countries should contain one benchmark station to
provide a continuing series of consistent observations on hydrological and related climatological
variables. Hydrological benchmark stations should
be established in areas that are relatively uninfl uenced by past or future anthropogenic changes.
Since long records are the essence of a benchmark
station, consideration should be given to existing
stations if they meet the other requirements. The
Reference Hydrometric Basin Network of Canada is
one such example (Harvey and others, 1999).
Climatological benchmark stations are known as
reference stations.


2.4.1.4.3 Representative basins
A representative basin is desirable in each natural
region – especially in those regions where great economic growth is expected or where the hydrological problems are particularly difficult. In their simplest form, they permit the simultaneous
study of precipitation and runoff, thus helping to
make up for deficiencies in short periods of
observation and low densities of minimum
networks.


2.4.1.4.4 Project stations
These are stations established for a limited span of
time, for specific purposes, often research oriented.
Other frequent objectives may be investigations
before or after physical interventions in the
catchment, or for supplementing the regional
coverage of the basic network. Project stations are
characterized by:
(a) Limited lifetime;
(b) Data quality depending on purpose.
2.4.1.5 Conducting a network analysis
Figure I.2.6 lays out the steps that should be taken
in conducting a review and redesign of an existing
hydrological network. Such reviews should be
conducted periodically to take advantage of the
reduction in hydrological uncertainty brought
about by the added data since the last network analysis and to tune the network to any changes in the socio-economic environment that may have transpired. The steps of the analysis are discussed
individually below.

Institutional set-up
The roles and aims of all of the organizations
involved in various aspects of water resources
management should be defined and identified,
particularly legislative responsibilities.
Communication links between these organizations
should be improved to ensure coordination and
integration of data-collection networks.

Purposes of the network
The purposes of the network in terms of the users
and uses of the data should be identifi ed. Data
users and uses can vary temporally and spatially.
There is also a need to identify potential future
needs and incorporate these into the design as
well.


Objectives of the network
Based on the purpose of the network, an objective
or set of objectives can be established in terms of
the information required. An indication of the
consequences of not being able to provide this
information may prove useful later.


Establish priorities
If there is more than one objective, priorities need
to be set for later evaluation. If all objectives can be
met within the budget, then this is not needed.
However, if they cannot be met, then the lowerpriority objectives may not be met fully.


Assess existing networks
Information on the existing networks should be
compiled and interpreted to determine if the
current networks fulfi l the objectives. This may
include comparisons with other basins and/or
networks.


Network design
Depending on the available information and the
objectives defined, the most appropriate network design technique or techniques should be applied.

This may be simple hydrological characteristics,
regression relationships, or more complex
network analysis using generalized least squares
methods.

Optimize operations
Operational procedures account for a signifi cant
portion of the cost of data collection. This includes
the types of instrument, frequency of station visits
and structure of field trips. The minimum-cost
operational procedures should be adopted.


Budget
Based on the identified network and operational
procedures, the cost of the operation of the network
can be established. If this is within the budget, the
next step can be followed. If not, either additional
funding must be obtained or the objectives and/or
priorities need to be examined to determine where
costs may be reduced. The process adopted should
allow the designer to express the impact of insuffi –
cient funding in terms of not meeting objectives or
reduced information and net impacts.


Implementation
The redesigned network needs to be implemented
in a planned manner. This will include both shortand long-term planning horizons.


Review
Since a number of the above components are
variable in time, a review can be required at the
instigation of any particular component – for
example, changes in users or uses, or changes in the
budget. To be ready to meet such changes, a
continuous review process is essential.


2.4.2 Density of stations for a network
The concept of network density is intended to serve
as a general guideline if specific guidance is lacking.
As such, the design densities must be adjusted to
refl ect actual socio-economic and physio-climatic
conditions. Computer-based mathematical analysis
techniques should also be applied, where data are
available, to optimize the network density required
to satisfy specific needs.


As stated in 2.4.1.3.1, the minimum network is one
that will avoid serious deficiencies in developing
and managing water resources on a scale commensurate with the overall level of economic
development and environmental needs of the
country. It should be developed as rapidly as possible, incorporating existing stations, as appropriate.
In other words, such a network will provide the
framework for expansion to meet the information
needs of specifi c water uses.


In the following sections, minimum densities of
various types of hydrological stations are
recommended for different climatic and
geographic zones. These recommendations are
based on the 1991 review of Members’ responses
regarding the WMO basic network assessment
project (WMO/TD-No. 671) and are presented in
Table I.2.6. However, these recommended network
densities are being revisited through a study
undertaken by the Commission for Hydrology and
the revised recommended densities will be placed
on the Website as part of the electronic version of
the Guide.


It is impossible to define a sufficient number of
zones to represent the complete variety of hydrological conditions. The simplest and most precise
criterion for the classification of zones would be on
the basis of the areal and seasonal variation of rainfall. Each country could present a good map of
annual precipitation and a minimum network
could be developed from this.

However, this would not help countries that need a network most as
they have very few prior records, and the establishment of a good precipitation map is not possible.
Also, the countries with very irregular rainfall distribution need to be considered as a special category.
In such cases, it is not advisable to base the classifi –
cation on this one characteristic.


Population density also affects network design. It is
almost impossible to install and operate, in a satisfactory manner, a number of stations where
population is sparse unless the stations are highly
automated. Sparsely settled zones, in general, coincide with various climatic extremes, such as arid
regions, polar regions or tropical forests.
At the other extreme, densely-populated urban
areas need a very dense raingauge network for both
temporal and spatial resolution of storms and for
design, management and real-time control of the
storm-drainage systems and for other engineering
applications.


From these considerations, a limited number of
larger zones have been defi ned for the defi nition of
density norms in a somewhat arbitrary manner
adopting some general rules.

Six types of physiographic regions have been defined for minimum
networks:


(a) Coastal;
(b) Mountainous;
(c) Interior plains;
(d) Hilly/undulating;
(e) Small islands (surface areas less than 500 km2);
(f) Polar/arid.
For the last type of region, it is necessary to group
the areas in which it does not seem currently possible to achieve completely acceptable densities
because of sparse population, poor development of
communications facilities, or for other economic
reasons.


2.4.2.1 Climatological stations
The following types of data are collected at a climatological station in the basic network: precipitation,
snow survey and evaporation. It is understood here
that evaporation or snow-measuring stations, particularly the former, will generally measure temperature,
humidity and wind because these meteorological
elements affect evaporation and melting.


2.4.2.1.1 Precipitation stations
If one follows certain principles of installation and
use, the small number of stations in the minimum
network can furnish the most immediate needs. In
general, precipitation gauges should be as uniformly
distributed as is consistent with practical needs for
data and the location of volunteer observers. In
mountainous regions, attention must be given to
vertical zonality by using storage gauges to measure
precipitation at high altitudes. Precipitation gauges
may be designed specifically to measure snow-water
equivalent, either through the addition of shielding to reduce under-catch due to wind or through
the use of pressure sensors. Periodic manual snow
surveys may be used to supplement the network,
but they should not be counted as part of the
network.


The network should consist of three kinds of
gauge:
(a) Standard gauges – These gauges are read daily
for quantity. In addition to daily depth of
precipitation, observations of snowfall, the
depth of snow on the ground and the state of
the weather are to be made at each standard
precipitation station;
(b) Recorders – In developing networks, it is advisable to aim to have at least 10 per cent of
such stations. The greatest density of recording stations should be achieved in those areas
subject to intense, short-duration rainfalls.
Such stations will provide valuable information
on the intensity, distribution, and duration of
precipitation.
For urban areas where the time resolution
needed for rainfall measurements is of the order
of one to two minutes, special attention should
be paid to the time synchronization of the
raingauges.

For reliable measurements, tipping bucket raingauges with an electronic memory
(or another computer readable medium) are
recommended.
In assigning priorities to locations for recording-raingauge installations, the following types
of areas should be given priority: urban areas
(population in excess of 10 000) where extensive drainage systems are likely to be constructed,
river basins in which major river control
systems are anticipated or are in operation,
large areas inadequately covered by the existing network and special research projects;
(c) Storage gauges (totalizers) – In sparsely settled
or remote regions, such as in desert or mountainous terrain, storage gauges may be used.
These gauges are read monthly, seasonally, or
whenever it is possible to inspect the stations.
Location of precipitation gauges relative to streamgauging stations –

To ensure that precipitation data
are available for extending streamflow records, flood-forecasting purposes or hydrological analysis,
coordination of the locations of the precipitation
gauges with respect to those of the stream gauges is
of great importance. Precipitation gauges should be
located so that basin precipitation can be estimated
for each stream-gauging station. These will usually
be located at or near the stream gauge and in the
upper part of the gauged drainage basin. A precipitation gauge should be located at the site of the
stream gauge only if the observations will be representative of the general area. There can be cases in
which it is desirable to locate the precipitation
gauge some distance away from the stream gauge,
as for instance when the stream gauge is in a narrow,
deep valley.


2.4.2.1.2 Snow surveys
Where applicable, observations of snowfall, water
equivalent of snow and depth of snow on the
ground should be made at all precipitation stations
in the minimum network.
The water equivalent of snow at the time of maximum accumulation is an indication of total seasonal
precipitation in regions where winter thaws and
winter snow melt are insignificant. In such regions,
surveys of the snow cover on selected courses may be
useful in estimating seasonal precipitation at points
where the normal observations are unavailable. Such
snow-cover surveys will also provide useful information for river forecasting and flood studies.
Snow-cover surveys are conducted by personnel
equipped for sampling the accumulated snow and
for determining its depth and water equivalent
(3.5). The number of snow courses and their location and length will depend upon the topography
of the catchments and the purposes for which the
data are being collected. The full range of elevation
and the types of exposure and vegetation cover in
the area of interest should be considered in selecting representative courses. It is suggested that one
course for 2 000 to 3 000 km2 is a reasonably good
density for less homogeneous regions, and one
course for 5 000 km2 in homogeneous and plain
areas. However, each case must be considered on its
own merits, and these generalities must not be
applied indiscriminately.
In the early stages of network development, snowcover surveys will usually be made only once a
year, near the expected time of maximum accumulation. It will be desirable, later on, to extend
the operation to include surveys at regular intervals throughout the snowfall season. As soon as it
becomes feasible, these periodic snow surveys
should be augmented by regular measurements of
snow precipitation and observations of related
meteorological factors, such as radiation, soil
temperature and wind velocity.


2.4.2.1.3 Evaporation stations
Evaporation can be estimated indirectly in the
water-budget, energy-budget and aerodynamic
approaches, by extrapolation from pan measurements or directly through use of eddy-correlation
equipment (Chapter 4). An evaporation station
consists of a pan of standard national design where
daily observations of evaporation are made, together
with daily observations of precipitation, maximum
and minimum water and air temperatures, wind
movement and relative humidity or dewpoint
temperature.


Evaporation plays an important role for long-term
studies of the water regime of lakes and reservoirs
and for water management. In such cases, the
number and distribution of evaporation stations
are determined according to the area and confi guration of the lakes and the climatic region or regions
involved.


2.4.2.2 Hydrometric stations
2.4.2.2.1 Streamflow stations
The main objective of the stream-gauging network
is to obtain information on the availability of
surface-water resources, their geographical distribution and their variability in time. Magnitude and
frequency of fl oods and droughts are of particular
importance in this regard.
In general, a sufficient number of streamflow
stations should be established along the main stems
of large streams to permit interpolation of discharge
between the stations. The specifi c location of these
stations should be governed by topographic and
climatic considerations. If the difference in fl ow
between two points on the same river is not greater
than the limit of error of measurement at the
station, then an additional station is unjustifi ed. In
this context, it must also be stressed that the
discharge of a small tributary cannot be determined
accurately by subtracting the fl ows at two main
stream-gauging stations that bracket the mouth of
the tributary. Where the tributary fl ow is of special
interest in such a case, a station on the tributary
will be required. It will usually take its place as a
secondary station in the minimum network. The
streamfl ow stations may be interspersed with stage
stations (2.4.2.2.2)

Wherever possible, the base stations should be
located on streams with natural regimes. Where
this is impractical, it may be necessary to establish
additional stations on canals or reservoirs to obtain
the necessary data to reconstruct the natural fl ows
at the base stations. Computed fl ows past hydroelectric plants or control dams may be useful for
this purpose, but provisions will have to be made
for calibration of the control structures and turbines
and for the periodic checking of such calibrations
during the life of the plants.


Stations should be located on the lower reaches of
the major rivers of the country, immediately above
the river mouths (usually above tidal infl uence) or
where the rivers cross borders. Stations should also
be located where rivers issue from mountains and
above the points of withdrawal for irrigation water.

Other hydrometric stations are situated at locations, such as where the discharge varies to a considerable extent, below the points of entry of the major tributaries, at the outlets from lakes, and
where large structures are likely to be built.
Hydrometric stations are often established at major
cities to meet a number of societal needs.
To ensure adequate sampling, there should be at
least as many gauging stations on small streams as
on the main streams. However, a sampling procedure
for small streams becomes necessary, as it is
impracticable to establish gauging stations on all of
them. The discharge of small rivers is strongly
influenced by local factors. In highly developed
regions, where even the smallest watercourses are
economically important, network deficiencies are
keenly felt even on streams draining areas as small
as 10 km2.


Stations should be installed to gauge the runoff in
different geologic and topographic environments.
Because runoff varies greatly with elevation in
mountains, the basic network stations must be
located in such a way that they can, more or less
evenly, serve all parts of a mountainous area, from
the foothills to the higher regions. Account should
be taken of the varying exposure of slopes, which is
of great signifi cance in rough terrain, and to land
cover, which may vary with exposure and other
factors. Similarly, consideration should be given to
stations in districts containing numerous lakes, the
infl uence of which can be determined only through
the installation of additional stations.


2.4.2.2.2 River stages
Stage (height of water surface) is observed at all
stream-gauging stations to determine discharge.
There are places where additional observations of
water level only is needed as part of a minimum
network:
(a) At all major cities along rivers, river stages are
used for fl ood forecasting, water supply and
transportation purposes;
(b) On major rivers, at points between stream-gauging stations, records of river stage may be used
for fl ood routing and forecasting purposes.


2.4.2.2.3 Lake and reservoir stages
Stage, temperature, surge, salinity, ice formation,
etc., should be observed at lake and reservoir
stations. Stations should be established on lakes
and reservoirs with surface areas greater than
100 km2. As in the case of rivers, the network should
sample some smaller lakes and reservoirs as well.
2.4.2.2.4 Sediment discharge and sedimentation
Sediment stations may be designed either to measure total sediment discharge to the ocean or to
measure the erosion, transport and deposition of
sediment within a country, basin, etc. In designing
a minimum network, emphasis should be placed
on erosion, transport and deposition of sediment
within a country. An optimum network would
contain a sediment station at the mouth of each
important river discharging into the sea.


Sediment transport by rivers is a major problem in
arid regions, particularly in those regions underlain
by friable soils and in mountainous regions where,
for engineering applications, the amount of sediment loads should be known.


The designer of a basic network must be forewarned
that sediment-transport data are much more expensive to collect than other hydrological records.
Consequently, great care must be exercised in selecting the number and location of sediment-transport
stations. Emphasis should be placed on those areas
where erosion is known to be severe. After a few
years of experience, it may be desirable to
discontinue sediment measurements at those
stations where sediment transport no longer appears
to be of importance.


Sediment-transport data may be supplemented by
surveys of sediment trapped in lakes or reservoirs.
Echo-sounding devices are useful for this purpose.
However, information obtained in this way is not
considered a substitute for sediment-transport
measurements at river stations. Sediment discharge
measurement and the computation of sediment
load are covered in 5.5.

2.4.2.2.5 Water-quality stations
The usefulness of a water supply depends, to a large
degree, on its chemical quality. Observations of
chemical quality, for the purposes of this Guide,
consist of periodic sampling of water at
stream-gauging stations and analyses of the
common chemical constituents. ISO Technical
Committee 147 has prepared over 200 international
standards pertaining to fi eld sampling for waterquality and analytical methods.


The number of sampling points in a river depends
on the hydrology and the water uses. The greater
the water-quality fluctuation, the greater the
frequency of measurement required. In humid
regions, where concentrations of dissolved
matter are low, fewer observations are needed
than in dry climates, where concentrations,
particularly of critical ions such as sodium, may
be high.


2.4.2.2.6 Water temperature
The temperature of water should be measured
and recorded each time a hydrometric station is
visited to measure discharge or to obtain a sample
of the water. The time of day of the measurement
should also be recorded. At stations where daily
stage observations are made, temperature observations should also be made daily. These
observations, the cost of which is negligible, may
provide data which are useful in studies of aquatic
life, pollution, ice formation, sources of cooling
water for industry, temperature effects on sediment transport, solubility of mineral constituents,
or climate change.


2.4.2.2.7 Ice cover on rivers and lakes
Regular observations of ice cover should include
the following:
(a) Visual observations of various processes of ice
formation and of ice destruction, with recording of date of fi rst occurrence of fl oating ice, date
of total cover, date of break-up of the ice, and
date at which the ice has vanished completely.
These observations should be made on a daily
basis;
(b) Simultaneous measurement of ice thickness at
two or three points near each selected hydrometric station should be made once every 5 to
10 days. The location of measurement points is
chosen from detailed surveys of ice cover made
at the beginning of the observing period of the
stations.


2.4.3 Specific requirements for water
quality


There are several approaches to water-quality monitoring. Monitoring can be accomplished through a
network of strategically located long-term stations,
by repeated short-term surveys, or by the most
common approach, a combination of the two. In
addition to the basic objectives of the programme,
the location of stations should take into account
the following factors:
(a) Existing water problems and conditions;
(b) Potential growth centres (industrial and
municipal);
(c) Population trends;
(d) Climate, geography and geology;
(e) Accessibility;
(f) Available human resources, funding, fi eld and
laboratory data handling facilities;
(g) Inter-jurisdictional considerations;
(h) Travel time to the laboratory (for deteriorating
samples);
(i) Safety of personnel.
The design of a sampling programme should be
tested and assessed during its initial phase to ensure
the effectiveness and effi ciency with respect to the
objectives of the study.


2.4.3.1 Water-quality parameters
The parameters that characterize water quality may
be classifi ed in several ways, including physical
properties (for example, temperature, electrical
conductivity, colour and turbidity), inorganic
chemical components (for example, dissolved
oxygen, chloride, alkalinity, fl uoride, phosphorous
and metals), organic chemicals (for example,
phenols, chlorinated hydrocarbons, polycyclic
aromatic hydrocarbons and pesticides), and biological components, both microbiological, such as
faecal coliforms, and macrobiotic, such as worms,
plankton and fi sh, which can indicate the ecological health of the aquatic environment.
A second classifi cation is done according to the
importance attached to the parameter. This will
vary with the type of water body, the intended use
of the water and the objectives of the monitoring
programme. Water-quality variables are sometimes
grouped into two categories:
(a) Basic variables (Table I.2.7) (UNEP, 2005);
(b) Use-related variables:
(i) Drinking water supplies;
(ii) Irrigation;
(iii) General quality for aquatic life.

A third classifi cation that is highly relevant
to sampling procedures is done according to
stability:
(a) Conservative (does not change materially with
time);
(b) Non-conservative (changes with time, but can
be stabilized for at least 24 hours by appropriate
treatment); or
(c) Non-conservative (changes rapidly with time
and cannot be stabilized).
The fi rst two groups can be measured by representative water samples subsequently analysed in the
laboratory. The third group needs to be measured
in situ.
2.4.3.2 Surface-water quality
Sometimes the programme objectives will precisely
defi ne the best locations for sampling in a river or
lake system. For example, in order to determine the
effect of an effl uent discharge on a receiving stream,
sampling locations upstream and downstream of
the discharge would be required. In other cases,
both location and frequency of sampling will be
determined by anti-pollution laws or by a requirement for a specifi c use of a water body. For example,
a permit to discharge surface waters may outline
details of monitoring, such as location, number of
samples, frequency and parameters to analyse.

Water-quality monitoring programmes may be supplemented by intensive, but infrequent, special purpose water-quality surveys aimed at understanding short-term fluctuations in water-quality
parameters. As well, special situations may call for
water-quality surveillance, the continuous, specifi c
measurement of selected parameters.
Sampling strategies vary for different kinds of water
bodies and media, for example, water, sediment, or
biota. Rivers mix completely within distances ranging from several kilometres to a few hundred
kilometres of any point source of pollution. Lakes
may be vertically stratifi ed because of temperature
or infl ows of high-density saline water. Groundwater
tends to fl ow very slowly, with no surface indication of the changes in its solutes taking place
below.


If the objective concerns the impact of human
activities on water quality in a given river basin, the
basin can be separated into natural and altered
regions. The latter can be further subdivided into
stationary zones for instance, over periods longer
than 10 years, and those in which the impact is
variable, such as agricultural, residential and industrial zones. In acid-deposition studies, an important
factor is the terrain sensitivity to the deposition.
Figures I.2.7 and I.2.8 provide some examples of
where and how sampling stations could be located
to meet specific objectives on river and lake
systems.

Table I.2.7. GEMS/Water basic variables


The next step in choosing sampling locations is to
collect relevant information about the region to be
monitored. The information sought includes
geological, hydrological and demographic aspects,
as well as the number of lakes and streams, size and
locations of aquifers, locations of existing waterquality or stream-gauging stations, flow rates,
climatic conditions in the catchment area, historical developments, present and potential municipal
and industrial centres, current water intakes and
waste outlets, natural salt springs, mine drainage,
irrigation schedules, fl ow regulation (dams), present
and planned water uses, stream or lake water quality objectives or standards, accessibility of
potential sampling sites (land ownership, roads and
airstrips), availability of services such as electricity,
and existing wate-quality data. Figure I.2.9 shows
the steps to be followed in selecting sampling sites.
The distance downstream to the point of complete
mixing is roughly proportional to the stream velocity and to the square of the width of the channel.

Figure I.2.7. Monitoring site: rivers


Rivers are usually sufficiently shallow that vertical
homogeneity is quickly attained below a source of
pollution. Lateral mixing is usually much more
slowly attained. Thus, wide swift-fl owing rivers may
not be completely mixed for many kilometres
downstream from the input point.

Figure I.2.9. Scheme for the selection of water quality sampling sites


Various protocols are recommended to determine
representative sampling in the cross-section of the
river, for example, six samples analysed in duplicate, at three positions across the river and two depths or
mid-depth samples at the quarter points, or other
equal distance points across the width of the river.
If a representative sample cannot be obtained, it is
advisable to select another site, either immediately
upstream or downstream.

The other alternative is to obtain a flow-weighted composite sample from samples collected on cross-section verticals.

Longitudinal mixing of irregular or cyclic discharges
into a river will have a secondary influence on the
location of a sampling site. Their effects need to be
taken into account in deciding the frequency of
sampling and interpreting data.

Sampling frequency depends on the purpose of
the network, the relative importance of the
sampling station, the range of measured values,
the time variability of the parameter of interest
and the availability of resources. In the absence
of suffi cient background information, an arbitrary frequency based on knowledge of local conditions is chosen.

After sufficient data have been collected, the frequency may be adjusted to reflect the observed variability. The frequency is
also influenced by the relative importance of the
station and whether or not the concentrations
approach critical levels for some substances
measured.


For lake stations, the recommended practice is to
sample fi ve consecutive days during the warmest
part of the year and five consecutive days every
quarter. Special cases include temperate-zone lakes
that experience stratification. These should be
sampled at least six times a year, together with the
occasional random sample, to cover the following
periods: during open water prior to summer stratification, during mixing following summer stratification, under ice, and during the periods of snow melt and runoff.

Similarly, additional samples
of rivers should be taken, if possible, after storm
events and during snow melt and runoff.


When parameters are plotted against time, some
cyclic variation may be apparent amidst the
random fluctuations. The detection of cyclic events
requires a sampling interval no longer than one
third of the shortest cycle time and sampling over
a period at least ten times longer than the time of
the longest cycle. Therefore, long-period cycles
will not be verified in the initial surveys, but
become apparent during the operation of the
network. In order to detect the cyclic variations,
some random sampling is desirable, for example,
on different days of the week or different hours of
the day.


2.4.3.3 Precipitation quality
In general, sampling sites should be selected to give
accurate and representative information concerning
the temporal and spatial variation of chemical
constituents of interest. Important factors to take
into consideration are prevalent wind trajectories,
sources for compounds of interest, frequency of
precipitation events (rain, snow, hail), and other
meteorological processes that influence the
deposition.

There are also local criteria to be considered:
(a) No moving sources of pollution, such as routine
air, ground, or water traffi c, should be within
1 000 m of the site;
(b) No surface storage of agricultural products,
fuels, or other foreign materials should be
within 1 000 m of the site;
(c) Samplers should be installed over fl at undisturbed land, preferably grass-covered,
surrounded by trees at distances greater than
5 m from the sampler. There should be no
wind-activated sources of pollution nearby,
such as cultivated fi elds or unpaved roads.
Zones of strong vertical eddy currents,
eddy zones leeward of a ridge, tops of windswept ridges and roofs of buildings, particularly, should be avoided because of strong
turbulence;
(d) No object taller than the sampler should be
within 5 m of the site;
(e) No object should be closer to the sampler
than a distance of 2.5 times the height by
which the object extends above the sampler.
Particular attention must be given to overhead wires;
(f) The collector intake should be located at least
1 m above the height of existing ground cover
to minimize coarse materials or splashes from
being blown into it;
(g) Automatic samplers require power to operate
lids and sensors, and in some cases for refrigeration in the summer and thawing in the winter.
If power lines are used, they must not be overhead. If generators are used, the exhaust must
be located well away and downwind from the
collector;
(h) To address issues on a continental scale, sites
should preferably be rural and remote, with no
continuous sources of pollution within 50 km
in the direction of the prevalent wind direction
and 30 km in all other directions.


It may not be possible to meet all of these criteria in
all cases. The station description should refer to
these criteria and indicate the exact characteristics
of each location chosen as a sampling site.

In the case of large lakes, the precipitation over the
lake may not be as heavy as along the shores and
the proportion of large particles may be smaller. In
order to sample in the middle of a lake, the sampler
can be mounted on a buoy, rock, shoal or small
island.


Event sampling is the preferred method for sampling
precipitation. Each rain shower, storm or snowfall
constitutes an event. The analysis of event-precipitation samples enables pollutants associated with a
particular storm to be determined, and a windtrajectory analysis can determine probable sources.
However, this sampling regime is very sensitive.
The same statistical considerations concerning
frequency of sampling apply here as for surfacewater sampling.

2.4.3.4 Sediment quality
Most of the selection criteria outlined in previous
sections also apply to sampling for sediment.
Therefore only additional special recommendations
will be described here.


For rivers where sediment-transport data are
required, it is necessary to locate the sampling sites
near a water quantity gauging station so that accurate stream discharge information is available at all
times. Sampling locations immediately upstream
from confluences should be avoided because they
may be subjected to backwater phenomena. In
streams too deep to wade, locate sampling sites
under bridges or cableways. When sampling from
bridges, the upstream side is normally preferred.
Sampling in areas of high turbulence, such as near
piers, is often unrepresentative. Attention also must
be paid to the accumulation of debris or trash on
the piers, as this can seriously distort the flow and
hence the sediment distribution. An integrated
sample obtained by mixing water from several
points in the water column according to their average sediment load can be considered as a
representative sample as long as there is a good lateral
mixing.


The best places to sample bottom deposits in fastfl owing rivers are in shoals, at channel bends and at
mid-channel bars or other sheltered areas where the
water velocity is at its minimum.


Sampling sites should be accessible during fl oods,
since sediment-transport rates are high during
these times.


For identification of peak pollution loads in rivers,
two cases must be considered:
(a) For pollution from point sources, sampling
should be done during low-fl ow periods, when
pollution inputs are less diluted;
(b) When pollutants originate from diffuse sources
such as runoff from the land of agricultural
nutrients or pesticides, sampling must be
focused on fl ood periods during which the
pollutant is washed out of the soil.
If one of the objectives is to quantify the transport
of sediment in the river system, it should be noted
that peak concentrations of sediment do not necessarily correspond with times of peak fl ow. Also, a
series of high fl ow rates will lead to progressively
lower sediment peaks – an exhaustion effect arising
from the depletion of material available for
re-suspension.


For lakes, the basic sampling site should be located
at the geographic centre of the lake. If the lake is
very large (area > 500 km2), several base stations
may be needed. If various sediment types must be
sampled, then data from acoustic surveys (echosounders) can be used both to identify the type of
surfi cial material (sand, gravel or mud) and to indicate the presence of layering below the surface.


Secondary sampling sites should be located between
the base station and major tributary inlets or pollutant sources. A common strategy is to place points
down the long axis of the lake with occasional
cross-lines. Three to fi ve stations should usually
give a good approximation to the sediment quality
of an average size lake. For statistical validity,
however, a larger number of sampling sites will
probably be required.


Sampling frequency in lakes is affected by the generally low concentrations of suspended sediment.
Sediment traps should be operated during periods
of maximum and minimum algal productivity and
at times of high input of sediment from rivers.
Repeat sampling of bottom sediments in lakes
needs to take into account the rates of sediment
accumulation. Basins in cool temperate climates
often have accumulation rates in the order of
0.1–0.2 mm per year. A resampling period of fi ve
years would then be too soon to provide worthwhile new information, unless the presence of a
new pollutant is to be tested.


2.4.3.5 Groundwater quality
A great deal of hydrogeological information may be
necessary to plan the sampling strategy for aquifers.
Water levels, hydraulic gradients, velocity and
direction of water movements should be known.

An inventory of wells, boreholes and springs fed by
the aquifer should be drawn up, and details of land
use should be recorded.


Groundwater samples are taken from drainage
water, open wells and drilled wells. Wells should be
sampled only after they have been pumped long
enough to ensure that a fresh sample has been
obtained. This is particularly necessary where a well
has a lining subject to corrosion.


An existing well is a low-cost choice, although wells
are not always at the best location or made of noncontaminating materials. A well that is still in use
and pumped occasionally is preferable to one that
has been abandoned. Abandoned or unused wells
are often in poor condition with damaged or leaky
casings and corroded pumping equipment. It is
often diffi cult to measure their water levels, and
they may be safety hazards.
Changes in groundwater quality can be very slow
and are often adequately described by monthly,
seasonal or even annual sampling schedules.


2.4.4 Operational data acquisition
networks
Many types of hydrological forecasts are compiled
on the basis of data from networks. Information
may include measurements, as well as details of
the operation of water-management and fl oodprotection works. A forecast system should make
use of data from the basic network (2.4.1.3) as far
as possible. The scope of the forecast network is
determined by:
(a) User demands for forecasts at specifi ed locations and for current information on the status
of water bodies;
(b) The network density needed to describe the
hydrological characteristics and the dimensions of water bodies;
(c) The technology for data transmission to the
forecast centre;
(d) The representativeness of the observations;
(e) The media for issuing forecasts.
The information on water-management operations
should be organized to fi t in with the normal operational routines of the water-management agencies
that supply the information.


A schedule of reports transmitted to the forecast
centre by non-automatic monitoring stations
should be drawn up, and the reports should be
classified according to whether they are regularly or
occasionally transmitted. The regular reports should
include daily information on water levels, discharge
and temperature and, where appropriate, ice
phenomena, as well as observations every 5 or
10 days on ice thickness, snow depth and water
equivalent. The occasional reports contain
emergency information on significant changes in
the regime of water bodies and operational control
strategies, as well as specially requested reports that
are needed to define the development of particular
hydrological phenomena.


The Casebook on Hydrological Network Design Practice
(WMO-No. 324) gives examples of spatial densities
for various hydrological variables and the general
principles for determining them based on the time
and space variability.


2.4.5 Network-strategy options
In addition to seeking to improve representativeness of existing surface-water data networks,
Hydrological Services should develop more
comprehensive monitoring strategies. For selected
basins, the hydrometric data-collection activities
need to be integrated with sediment, water quality, meteorology and aquatic-habitat programmes


(2.4.1.4). For example, concerns for sediment-associated contaminant transport require knowledge
of the source, pathways and fate of fi ne particles.
This requires an understanding of both the fl ow
and sediment regimes. Whether for the interpretation of concentrations or for calculating
contaminant loadings, such integrated monitoring requires close coordination at all stages from
planning to reporting.


Integrated planning of related data networks should
be developed to maximize the effectiveness of all
water-data programmes. Significant efforts are
required to defi ne network needs from many different
perspectives, and, ultimately, to coordinate the data
collected on a watershed basis so that adequate water
data, that is, precipitation, runoff, groundwater and
water quality, are available to meet future needs.
Present monitoring programmes can be enhanced
by the use of supplementary studies. For example,
river studies of sediment sources and morphologic
change (Church and others, 1989; Carson, 1987)
supplement regular programme data to determine
the river behaviour. This knowledge, which is not
acquired from monitoring studies alone, is being
used for fisheries management, river-engineering
studies and water-quality studies.


On a different scale, water-quality considerations
are increasingly important to urban drainage design.

The design of appropriate monitoring programmes
should consider short-interval sampling, integrated
precipitation and runoff monitoring, and extremely
rapid response times if the data are to be useful.
These conditions are quite different from those
covered by standard monitoring procedures. The
use of computer models is an additional strategy for
enhancing the information derived from watermonitoring activities. In certain circumstances,
monitoring-network designs can be improved by
the use of models.


2.5 DATA COLLECTION
2.5.1 Site selection
Once the network design phase has been completed,
the operational requirements have established the
general location of the data-collection sites, and the
types of instrumentation have been identifi ed, the
best specifi c site in the general location is selected
to meet the requirements of the instrumentation as
outlined in subsequent chapters of this Volume
(5.3.2.1 and 5.4.2). Modifi cations to the site may be
necessary to ensure the quality of the data, for
example, clearing and control stabilization.


When a site has been selected and the instrumentation has been installed, two types of data will be
collected at the site: descriptive details of the site
and its location, and the hydrological observations
that it has been established to measure. Once established, the installation should be operated and
maintained to its predetermined standard. In
general, this involves the execution of an adequate
schedule of inspection and maintenance to ensure
continuity and reliability of data, and the development of routine check measurements and
calibrations to ensure data of the required
accuracy.


2.5.2 Station identifi cation
Two aspects should be considered to ensure the
historical documentation of details of a data collection site: the institution of an identification
system and the archival of descriptive information.

2.5.2.1 Identification of data-collection sites
Every permanent site should be given a unique
identifier that will be used to denote all data and
other information pertinent to the site. Such identifiers are usually numeric, but they may also be
alphanumeric.


Frequently, more than one service or agency may be
operating data-collection sites in one particular
region or country. The acceptance by all parties of a
single, unique system of site identification will
facilitate data interchange and the multiparty
coordination of data-collection activities. The region
chosen should be determined by drainage basin(s)
or climatic zones, and part of a site’s identification
should reflect its location in the region.


The site identification can be simply an accession
number, that is, a sequential number assigned as
stations are established. For example, site identify –
cation in the Canadian National Water Quality
Data Bank, NAQUADAT, represents a sophisticated
system designed for computer processing. It has a
12-digit alphanumeric code, which is the key
element in storing and retrieving data in the computer system. This number is composed of several
subfi elds (UNEP/WHO, 1996), as follows:


(a) Type of water – a two-digit numerical code indicating the type of water sampled at any given
location, such as streams, rivers and lakes, or
precipitation. The meaning of this code has
been extended to include other types of aquatic
media. A list of all currently assigned codes is
given in Table I.2.8;
(b) Province, basin and sub-basin – three pairs
of digits and letters identifying the province,
basin and sub-basin;
(c) Sequential – a four-digit number assigned
usually by a regional offi ce.
For example, station number 00BC08NA0001 indicates that the sampling site is on a stream, in the
province of British Columbia, in basin 08 and in
sub-basin NA, and the sequence number is 1. Station
number 01ON02IE0009 is on a lake, in the province of Ontario, in basin 02 and in sub-basin IE and
the sequence number is 9.
WMO has accepted a coding system for station
identification (Moss and Tasker, 1991) that is similar to (b) and (c) of the NAQUADAT system.


Another well-known coding system for sampling
points is the River Mile Index used by the
Environmental Protection Agency of the United
States as part of the STORET system. In this system,
the location of a sampling point is defined by its
distance and hydrological relationship to the
mouth of a river system. It includes major and
minor basin codes, terminal stream numbers, the
direction and level of streamflow, the mileages
between and to confluences in the river system,
and a code to identify the stream level on which
the point is located.

Table I.2.8. NAQUADAT codes for types of aquatic media

2.5.2.2 Descriptive information
In many instances the value of the data will be
enhanced if the user can relate it to the details of
the history of its collection as part of the routine
production of metadata. To this end, a station registration file should record the details of each station.


The level of detail will of course vary with the
parameter monitored. Typical information would
include the station name and location details, the
station type, the associated stations, establishing/
operating/owner authorities, the elevation details,
the frequency of observation, the operating periods
and the details of installed equipment. Additional
items specific to the station type should also be
included. Selected information from this text file
should be attached routinely to any data output
(Chapter 10).

A historical operations file of more detailed information should also be prepared for release as required (Chapter 10). Again, the level of detail
will vary with the type of observations being
recorded.

A stream station may include details such as climate zone and rainfall and evaporation notes, geomorphology, landforms, vegetation,
land use and clearing, and station details. Typical
components of such a file would include the
station description, a detailed sketch of the site, a
map showing the location of the site in the region,
and a narrative description of the site and region.
Some examples of the format of such fi les can be
found in the UNEP (2005) and Environment
Canada (1983) publications. Figure I.2.10 is an
example of one format.


2.5.2.2.1 Station description
An accurate description of the sampling location
includes distances to specifi c reference points. It is
important that these reference points be permanent and clearly identifi ed. For example, “5 metres
north-west of the willow sapling” is a poor designation for a data site. An example of a useful
description is “30 metres downstream from Lady Aberdeen Bridge (Highway 148), between Hull and
Pointe Gatineau and 15 metres off the pier on the
left side looking downstream”. If hand-held global
positioning devices are available, the geographic
coordinates of the sampling location should be
determined and recorded on the station description. The dates that the station was first established
and that data collection was commenced should
also be recorded.


For streamflow and water-quality data stations,
location information should also include descriptions of the water body above and below the
station. These should include water depths, a
description of the banks on either side of the water
body and the bed material. A description of the
water body should include any irregularities in
morphology that might affect the flow of water or
its quality. Such irregularities may include a bend
in a river, a widening or narrowing of the channel,
the presence of an island, rapids or falls, or the
entry of a tributary near the station. A description
of the banks should mention slope, bank material
and extent of vegetation.

Bed or sediment material may be described as rocky, muddy, sandy, vegetation-covered, etc. Station-location descriptions should mention seasonal changes that may hinder year-round data collection. Additional information in the case of lakes could include surface area,
maximum depth, mean depth, volume and water
residence time.


Additional information about conditions, either
natural or man-made, which may have a bearing
on the data should be recorded. Past and anticipated land disturbances and pollution sources
should be mentioned, for example, forest fi res, road
construction, old mine workings, and existing and
anticipated land use.


2.5.2.2.2 Detailed sketch of station location
A sketch of the location and layout of the station
(including distances expressed in suitable units)
with respect to local landmarks and permanent
reference points, such as benchmarks, should be
prepared (Figure I.2.11). Sampling or measuring
sites and equipment locations should be prominently shown on the sketch.


2.5.2.2.3 Map
A large-scale map (Figure I.2.12) that locates the site
with respect to roads, highways and towns should
be included. The combination of the map and the
sketch of the station location should provide
complete location information. An investigator
travelling to the site for the first time should have
enough information to locate the station confi –
dently and accurately.


2.5.2.2.4 Coordinates
Geographical coordinates are recorded as latitude
and longitude and, in addition, coordinates may
be recorded in other reference systems such as universal transverse mercator (UTM) coordinates or
legal land descriptions. If the site is on a stream, its
distance upstream from a reference point, such as
reference station or a river mouth should be
recorded. National grid references, if available,
should also be provided. For the international
GLOWDAT (that is, GEMS/WATER data bank (UNEP,
2005) station), one entry is the WMO code for the
octant of the globe for the northern hemisphere:
0, 1, 2 and 3 for 0–90°W, 90–180°W, 180–90°E
and 90–0°E, respectively (WMO-No. 683).
Correspondingly, for the southern hemisphere the
codes are: 5, 6, 7 and 8 for 0–90°W, 90–180°W,
180–90°E and 90–0°E (WMO-No. 559).


Latitude and longitude values should be obtained
using a global positioning system or, if that is not
possible, from 1:50 000 or 1:250 000 topographical
maps. Points on a 1:250 000 map can be located to
about ±200 m and on a 1:50 000 scale to about
±40 m (WMO-No. 559). If available, navigational
charts can be used to provide more accurate values
than the topographical maps.


2.5.2.2.5 Narrative description
For streamflow and water-quality sites, it is recommended that the narrative description begin with
the name of the river, stream, lake, or reservoir,
followed by its location (for example, upstream or
downstream) and its distance (to 0.1 km or better)
from the nearest town, city, important bridges,
highways or other fi xed landmarks. The name of
the province, territory or other geopolitical division
should also be included.


Information concerning changes at the site,
including instrumentation changes, should be
added to the narrative description to provide a
historical description of the site and the region that it represents.

Chapter 10 contains a suggested
format for such information.

2.5.3 Frequency and timing of station
visits
The frequency and timing of readings and thus
visits to the site should be determined by the anticipated data usage and should be adequate to define
the observations over time. Station visits will thus
be for purposes of observation or collection of data
and for maintenance of the site.


When the variable of interest at the site is changing
rapidly, visits to manual stations must be more
frequent if a valid record is to be maintained. Under
such conditions, it may be more efficient to install
automatic recording equipment or real-time transmission if funds and trained staff are available. This
applies particularly where more frequent observations are desirable for hydrological purposes during
storms and flood periods, as well as in tidal reaches
of rivers.


2.5.3.1 Manual stations
There is considerable merit in encouraging the
taking of observations at climatological stations at
specified synoptic hours. WMO recommends
(WMO-No. 544) that the time at which three-hourly
and six-hourly weather observations are taken at
synoptic stations are 0000, 0300, 0600, 0900, 1200,
1500, 1800 and 2100 universal time coordinated
(UTC). In most countries, such stations are the key
stations of the meteorological and climatological
observation programmes. If the observer is to take
three observations per day, the synoptic hours most
conveniently related to normal times of rising and
retiring and that nearest noon should be specifi ed.
For stations at which only one or two observations
per day are taken, it should be possible to select
synoptic hours for the observations.


It is recommended that all observers making only
one observation per day should have a common
observation time, preferably in the morning.
Some streams, for example small mountain-fed
streams, may experience diurnal fl uctuations in
water levels during some seasons. Stage observations should initially be made several times a day at
new stations to ensure that a single reading is an
adequate representation of daily water level. Also,
small streams may exhibit “fl ashy” behaviour in
response to rain storms. Additional stage readings
should be obtained during these times to adequately
defi ne the hydrograph. Stage observations should
also be made at the time of water-quality
sampling.


While it is desirable to have regular observations
at synoptic hours, in some cases this will not be
possible. In these cases, it is important that observations be taken at the same time each day and
that this time be recorded in UTC or local standard time using 24-hour clock designations. If
“summer time” (daylight saving time) is introduced for part of the year, arrangements should
be made to have observations taken at the same
hour, by UTC, as in the period prior to and following “summer time”.
The designated time of climatological observations
should be the end of the time at which the set of
observations is taken at a station. The set of observations should be taken, if possible, within the
10-minute period prior to the stated observational
time. However, it is important that the actual time
of observation be recorded carefully, whether the
observation is taken at a standard time or not. In
tidal reaches of rivers, the times of observation
should be related to the tidal cycle.


2.5.3.2 Recording stations
The frequency and timing of visits to recording
stations will be constrained by the length of time
that the station can be expected to function without
maintenance. For example, some continuous
rainfall recorders record on a weekly strip chart and,
thus, require weekly visits to remove and replace
charts. Other instruments have much larger data
storage capabilities and, therefore, require less
frequent visits.

A balance must be achieved between
the frequency of the visits and the resultant quality
of the data collected. Too long a time between visits
may result in frequent recorder malfunction and,
thus, in loss of data, while frequent visits are both
time consuming and costly. Various studies have
been carried out on the cost-effectiveness and
efficiency of data collection. Further details are
found in the Proceedings of the Technical Conference
on the Economic and Social Benefi ts of Meteorological
and Hydrological Services (WMO-No. 733).


The frequency of the visits may also be determined
by accuracy requirements of the data. Some data collection devices may suffer a drift in the relationship between the variable that is recorded and that which the recorded value represents. An
example of this is a non-stable stage-discharge relationship. In such cases, visits to the station are required periodically in order to recalibrate the equipment or the measurement equations.

Figure I.2.10. Station-location forms

2.5.3.3 New technologies
The introduction of data loggers and telephone/
satellite data transmission may have a significant
impact on station inspection/data-collection
frequencies (2.5.6). However, it should be noted
that in order to ensure the quality of the data, regular station maintenance is necessary.


2.5.4 Maintenance of sites
The following maintenance activities should be
conducted at data-collection sites at intervals determined to ensure that the quality of the data being
recorded is adequate. These activities could be
conducted by the observer responsible for the sites,
if there is one. However, they should occasionally
be performed by an inspector (9.8.4).


All collection sites:
(a) Service the instruments;
(b) Replace or upgrade instruments, as required;
(c) Retrieve or record observations;
(d) Perform the recommended checks on retrieved
records;
(e) Carry out general checks of all equipment, for
example, transmission lines;
(f) Check and maintain the site to the recommended specifi cations;
(g) Check and maintain access to the station;
(h) Record, in note form, all of the above activities;
(i) Comment on changes in land use or vegetation;
(j) Clear debris and overgrowth from all parts of
the installation.


Streamflow collection sites:
(a) Check the bank stability, as necessary;
(b) Check the level and condition of gauge boards,
as necessary;
(c) Check and service the fl ow-measuring devices
(cableways, etc.), as necessary;
(d) Check and repair control structures, as necessary;
(e) Regularly survey cross-sections and take photographs of major station changes after events or
with vegetation or land-use changes;
(f) Record, in note form, all of the above activities
and their results;
(g) Inspect the area around or upstream of the site,
and record any signifi cant land-use or other
changes in related hydrological characteristics,
such as ice.
Further details are found in the Manual on Stream
Gauging (WMO-No. 519).
Flood gauging cannot be programmed as part of a
routine inspection trip because of the unpredictable
nature of floods. A flood action plan should be
established prior to the beginning of the storm
season and should include priority sites and types
of data required. If flood gaugings are required at a
site, the preparations must be made during the
preceding dry season so that all is ready during the
annual flood season. Additional measures may be
required if severe flooding is likely.

Preparations include:
(a) Upgrade site access (helipad, if necessary);
(b) Equip a temporary campsite with provisions;
(c) Store and check gauging equipment;
(d) Flood-proof instrumentation such as stage
recorders.
Following the recession of flood waters, particular
attention is required to ensure the safety and security of the data-collection site and to restore normal
operation of on-site instrumentation. In some cases
redesign and reconstruction of the site will be
required. This work should take into account information obtained as a result of the flood.


2.5.5 Observations
At all data-collection sites a value must fi rst be
sensed, then encoded or recorded, and finally
transmitted. Examples of the components of data
collection are displayed in Table I.2.9.


2.5.5.1 Manual stations
At the very minimum, observers should be equipped
with fi eld notebooks and/or station journals in
which the original observations are recorded as
they are taken. Forms should also be provided to
permit the observer to report observations daily,
weekly, fortnightly, or monthly, as required. The
fi eld notebook or station journal should be retained
by the observer in case the report is lost in transit.
The report forms should be designed to permit easy
copying of the results from the fi eld notebook or
station journal. A good approach is to have the
report form identical to a page in the notebook or
journal. At least, the various elements should be in
the same columns or rows in both. Space should be
allowed in the journal and, perhaps, in the report
form for any conversions or corrections that may
have to be applied to the original readings.

Alternatively, an observation notebook with carbon
paper between successive sheets will permit easy
preparation of an original form for dispatch to the
central office and a copy for the local station record.
This is not a satisfactory procedure where the

notebook is to be carried into the field as moisture
can easily make the entries illegible. The report
forms may also be coding forms suitable for direct
conversion to computer medium.


The value of data can be greatly enhanced – or
devalued – by the standard of the accompanying
documentation. Observers should be encouraged to
comment on any external influences that may
affect observations, whether they be related to
equipment, exposure, or short-term influence. In
addition, input formats and forms should be flexible enough both to allow comments to be appended
and for these comments to be accessible with the
final data. It is important that published comments
be expressed in standard terminology, and it is preferable that correct vocabulary be employed in the
field report.


There is also reason for setting up the processing
system so that quality coding or tagging is carried
out as the observations are made. This is particularly applicable to manual observations because it
encourages the making of judgements while the
conditions are being observed. Data from field
measurement books may be processed using optical
readers or portable fi eld computers that will allow
the direct input of observations into computer storage. Such devices allow for reduced data transfer
errors and automatic data quality checks.

Table I.2.9. The components of data collection


Field observations that may assist in interpreting
water quality should be entered on the report. These
observations may include unusual colours or odours
of the water, excessive algal growth, oil slicks,
surface films, or heavy fish kills. Such observations
may prompt the field investigator to take additional
observation-based samples, in addition to those
required by the routine schedule.

The types of samples and their preservation should be consistent with the types of analysis that the investigator
thinks is warranted by the prevailing conditions. If
additional samples are collected at sites other than
the established station, the description of their
locations should be recorded accurately.

2.5.5.2 Recording stations
At automatic recording stations, observations are
recorded in digital or graphical form. However, the
following observations should be recorded at the
time of any visits for data retrieval or station
maintenance:
(a) Site identifi cation number;
(b) Observations from independent sources at the
time of collection, for example, gauge boards
and storage rainfall gauges;
(c) Specifi c comments relating to the recording
device, including its status, current observation
and time.
Each inspection should be recorded by completing
a station-inspection sheet. Data may be recorded in
solid-state memory or perforated tape. Final extraction of observations from the recorded data may be
performed at computing facilities when removable
memory of perforated tape has been used as the
recording medium. However, portable computers
may be used to extract data directly from data
loggers and to verify the data before leaving the
station. Field verification allows any necessary
repairs or other changes to be made before leaving
the site.


Data loggers record data at specifi c time intervals
(as programmed by the user). Intelligent loggers
will also allow for data compaction and variability
of observation times. In the case of the observation
of multiparameters, the coordination of observations can also be performed by the intelligent field
logger. For example, rainfall data can be recorded at
a fi ve-minute interval or at every tip of a bucket, for
stage data when the level alters by more than 1 cm,
and water-quality parameters when stream height
alters by 10 cm and/or on a 24-hour basis.


With graphical recorders, observations are collected
continuously and processing of the data in the
office is required. Comments should be written on
the chart or noted on the inspection sheet if any
errors are detected. As with digital recorders, independent field observations should be made and
recorded during each site visit.


After a station has been in operation for a reasonable period, the frequency and timing of inspections
should be re-assessed in the light of the capabilities of the instrumentation and the requirements
for data at that site. In some cases, consideration
should be given to the real-time collection of data
via various communications options as a cheaper
method of data collection than regular site visits
(2.5.6).


2.5.5.3 Real-time reporting
There are many recording and non-recording
stations from which real-time data are required, for
example, in the operation of reservoirs, fl ood-warning and forecasting situations, and in some instances
as a cost-effective method of data collection.
Real-time data collected by fi eld observers must be
reported using a transmission facility, such as a
radio or the public telephone system, to the agency.
Similarly, recording stations must report via some
transmission facility. Recording devices may have
the advantage of being able both to transmit data at
prescribed intervals/parameter changes and be
interrogated by the collecting agency to determine
the current situation or reset observation intervals.
Data loggers may also provide information on the
current available storage capacity of the logger and
the condition of the available power supply.
Automated quality-control processes can be
developed in these situations.


2.5.5.4 Instructions for observers
Clearly written instructions must be provided to all
observers. These should contain guidance and
directions on the following matters:
(a) A brief description of instruments, with
diagrams;
(b) Routine care and maintenance of instruments
and actions to be taken in the event of serious
breakage or malfunctioning;
(c) Procedures for taking observations;
(d) Times of routine observations;
(e) Criteria for the beginning, ending and frequency
of special non-routine observations, for example, river-stage observations while water level is
above a predetermined height;
(f) Procedures for making time checks and putting
check observations on charts at stations with
recording instruments;
(g) Completion of fi eld notebooks or station
journals;
(h) Completion of report forms, including methods
of calculating means and totals with appropriate
examples;
(i) Sending of reports to the central offi ce;
(j) Special routines for real-time stations.
Such written instructions should be supplemented
by oral instructions by the inspector to the observer
at the time of installation of instruments and at
regular intervals thereafter.


The instructions should emphasize the importance
of regular observations with perhaps a brief account of how the observed data are used in water resources development, hydrological forecasting, or flood-control studies. Any special observations
that may be required during special periods, for
example, during fl oods, or any special reports that
are to be filed, should be specifi cally discussed.


Observers should be urged not to forget to fi ll in
the spaces for station names, dates and their signature. The necessity of reporting immediately any
instrument failure or signifi cant modifi cation of
the observing site should be emphasized.
Observers at stations equipped with automatic
recording instruments must be provided with
instructions on the method of verifying the
operation of digital recorders, changing charts and
taking check observations. These instructions must
stress the importance of annotating the chart with
all information that might be required for later
processing. This would include station identifi cation,
time on, time off, check-gauge readings and any
other entries that would make the record more
easily interpreted at a later time.


At stations with full-time personnel, the staff should
be suffi ciently well trained to abstract data from
recording instruments. For such stations, carefully
worded instructions on the method of abstracting
data and on the completion of report forms must
be provided. However, at many ordinary stations,
where observers may not be thoroughly trained, it
may be undesirable to require observers to undertake the relatively complex job of data abstraction.
In such cases, digital or graphical records should be
forwarded to a central office for processing of the
data.


2.5.6 Transmission systems
2.5.6.1 General
During recent years, the demands from users of
hydrological data have become more and more
complex; therefore, systems that include automatic
transmission of hydrological observations have
been incorporated into national networks. This has
also led to the need for developing codes to facilitate the formatting of observations for the
transmission and dissemination of forecasts.
Hydrological codes are discussed in 2.3.2. The
following describes different possibilities for transmission systems:
(a) Manual – The observer at the station mails
data or initiates radio or telephone calls to the
central offi ce on pre-arranged criteria;
(b) Manual/semi-automatic – The central offi ce
manually interrogates the remote automatic
station by telephone, Internet, radio or radio
telephone or satellite, and receives single
discrete values as often as interrogated. It is
possible to have automatic telephone-dialing
equipment in the central office that can make
calls in series;
(c) Automatic timed – Automatic equipment at
stations is programmed to initiate transmission
of a single, instantaneous observation and/or
past observations held in a storage register;
(d) Automatic event indicator – The station transmits automatically, by radio, telephone, Internet or satellite, a specifi ed unit of change of a
variable, for example, each centimetre change
in the stage of a river;
(e) Automatic – Data are transmitted by the station
and recorded at the central offi ce on a continuous basis.
2.5.6.2 Transmission links
The possible choices of transmission links include:
(a) Dedicated land-lines – These are used where
relatively short distances are involved and
commercial lines are not readily available;
(b) Commercial telephone and telegraph lines –
Telephone and telegraph systems can be used
whenever feasible. Equipment that permits
unattended reception of observations at the
central office is available. Measurements and
commands can be transmitted to and from the
remote site;
(c) Commercial cellular telephone networks – The
ever growing coverage of these networks,
together with better and more reliable equipment, make them an interesting and less
expensive option for moving data from a site
and into the central offi ce. The combination
of reliability and low cost makes it more realistic to collect data from stations with no realtime interest, from sites previously considered
as somewhat remote, to be transmitted using
commercial facilities. Cellular systems can be
used in the same way as standard telephone
lines and may continue to operate during an
extreme event when telephone lines fail;
(d) Direct radio links – These must be used when
requirements cannot be met by those facilities provided by landlines, or when distances or natural obstacles prevent the economic
installation of wires. Distances of several to
hundreds of kilometres may be spanned by
radio transmitters, depending upon the carrier
frequency and the transmitter power. At the
higher frequencies, the transmitter and receiver
must have a clear line-of-sight transmission
path. This limits the range without repeater stations to about 50 km. In all cases, the installation and operation of radio transmission
links are subject to national and international
regulations;
(e) Satellite links – Data transmission using satellites can take place in two ways: transmission
of data, as observed by sensors in the satellite
(such as imagery) or the use of the satellite to
relay data observed at remote ground stations
to central receiving locations. At the present
time, the science of observation and transmission or retransmission from satellites is developing rapidly. The data involved are available
either directly from the spacecraft or through
central data banks;
(f) The Internet – Internet Protocol communication in various forms, including the use of
mobile phone networks, makes this an interesting and less expensive way to send data,
especially if there is much data to transfer
or continuous transfer is wanted. Internet
communication works on a number of different physical communication paths, including
both mobile and ordinary telephone networks.
This makes it more reliable. In systems with a
large number of sites, it also makes the retrieval
time shorter and the communication system in
the main office much easier.


2.5.6.3 Factors affecting the choice of
transmission systems
When considering the possibility of including automatic transmission of data in any measuring system,
consideration should be given to the following:
(a) Speed with which data are required. This
depends upon the following factors:
(i) The speed with which changes in the
measured variable take place;
(ii) The time between the observation and
receipt of the data by conventional means,
versus automatic transmission systems;
(iii) The urgency of having this information
available for warnings or forecasts;
(iv) The benefits of forecasts from telemetered
data and economic losses due to lack or
delay of forecasts;
(v) The advantages of radio and satellite
transmission versus landlines in times of
storms and floods when these disasters
can destroy the more conventional means
of telecommunications at the time that
the information is most urgently needed;
(b) Accessibility of the measurement sites for quality control and maintenance;
(c) Reliability of the recording device. When local
climatic conditions are rigorous, the operation
of on-site mechanical equipment is difficult.
Under these situations, it may be more reliable to transmit data electronically to a central
climate-controlled offi ce. This system also
permits a continuous check of the operation of
the sensors;
(d) Staffi ng for operational, maintenance and logistic problems. It is important for these aspects to
be considered in the planning process and to
recognize that each individual project will have
its own particularities. Careful attention should
be given to the costs and benefi ts of all the
alternatives before any fi nal decision is made.
When designing a system for the automatic
transmission of data, the main components to
consider for staffi ng purposes are:
(i) Sensors and encoding equipment;
(ii) The transmission links;
(iii) Receiving and decoding equipment.
It is necessary to consider these components jointly
in the design stage. This is essential because the
special characteristics of any one component can
have serious consequences on decisions regarding
the others. If the ultimate use of the data transmission system is intended for forecasting, then
sensing, transmitting and receiving hydrometeorological data is an essential but an insufficient component of the forecast system. A forecast centre having personnel who are well-trained in preparing
forecasts and warnings, and in notifying persons at
risk is also fundamental (United Nations, 2004).
2.5.7 Water-quality monitoring
Chapter 7 provides details of instrumentation and
field practices for the collection of water-quality
data. The sampling locations, the sampling times,
the parameter identifications and the corresponding values must be recorded and coherence must
be maintained throughout the handling of the
data. If any one of these essential items is lacking,
then the whole effort is wasted.


2.5.7.1 Station identifi cation
The importance of an accurate written description
of each station location and the conditions under
which the samples are collected are discussed in
detail in 2.5.2.2.


2.5.7.2 Field sheets for water-quality
monitoring
Perhaps one of the most important steps in a
sampling programme is the recording on the field
sheets of observations, sampling date, time, location and the measurements made. All field records must
be completed before leaving a station. Additional
instructions are contained in 2.5.5.
Two examples of a systematic format for recording
fi eld analyses and observations are provided in
Figures I.2.13 and I.2.14. The formats shown in
these figures are appropriate for those personnel
that use computer systems for storing their results.


The format of Figure I.2.13 can be used by anyone
collecting water-quality data. Both formats can be
adapted to fit situations specific to a particular need.
The following information is usually recorded:
(a) Sampling site and date;
(b) Field-measured parameters;
(c) Instrument calibration;
(d) Sampling apparatus used and procedures;
(e) Quality control measures used;
(f) General remarks and fi eld observations.
2.5.7.3 Transportation of water-quality
samples
Once collected, some water samples must be transported to the laboratory. The mode of transportation
will depend on the geographic location and the
maximum permissible time lapse before analysis
for each constituent. The field investigator is
respons ible for delivering the samples to the
airline, bus, train or postal terminal on schedule so
that there will be minimal delay in sample transport. Logistics for sample transport and storage
should be determined before fieldwork is
initiated.

Figure I.2.13. Field sheet for use with NAQUADAT or similar computer system
Figure I.2.14. General format for a fi eld-sampling sheet

2.5.7.4 Field quality assurance in water quality monitoring
A field quality assurance programme is a systematic
process that, together with the laboratory and data storage quality assurance programmes, ensures a
specific degree of confidence in the data. A field
quality assurance programme involves a series of
steps. All equipment should be kept clean and in
good working condition, with records kept of calibrations and preventive maintenance. Standardized
and approved methodologies, such as those recommended in this Guide, should be used by field personnel.

The quality of data generated in a laboratory
depends on the integrity of the samples that arrive
at the laboratory. Consequently, the field investigator must take the necessary precautions to protect
samples from contamination and deterioration.
Further details on fi eld quality assurance are available in Chapter 7 of the present Guide; ISO Standards
(ISO 5667–14:1998 Water quality-Sampling –
Part 14: Guidance on quality assurance of
environmental water sampling and handling), in
the Water Quality Monitoring: A Practical Guide to the
Design and Implementation of Freshwater Quality
Studies and Monitoring Programmes (UNEP/WHO,
1996); and the Manual on Water Quality Monitoring:
Planning and Implementation of Sampling and Field
Testing (WMO, 1988).


2.5.8 Special data collection
2.5.8.1 Requirement
Data concerning severe storms and floods are very
important in determining design criteria for many
types of hydraulic structures. In general, regular
observation networks do not provide enough
detailed information on storm-rainfall distribution,
or on flood-peak discharges of tributary streams. In
addition, during severe floods, permanent streamgauge installations are sometimes overtopped or
washed away and the record is lost. For these
reasons, very valuable information can be obtained
by a fi eld survey crew in the area of a storm flood
immediately following a severe occurrence. In
addition, data from instruments, such as weather
radar, are often valuable in hydrological studies
(3.7).


2.5.8.2 Bucket surveys of storm rainfall
Measurements of rainfall from private, non-standard raingauges, and estimates that can be made
from various receptacles, such as pails, troughs and
barrels (provided these can be verified to have been

empty prior to the storm), can be used to augment
rainfall data from the regular observing network.

Eyewitness reports can be obtained of beginning
and ending times of rainfall and of periods of very
heavy rain. Care must be taken in interpretation of
bucket-survey data, and where discrepancies exist
between data from a bucket survey and the regular
observation network. Greater weight should usually
be given to the latter.
2.5.8.3 Weather-radar and satellite data
Data from weather radars and satellites are valuable
in determining the intensity and areal distribution
of rainfall and beginning and ending times of
precipitation over a specifi c river basin. For record
purposes, these data can be collected on photographic
fi lm or in digital form by a computer linked to the
radar. These digitized data can be readily transmitted
to forecast offices over computer networks.


2.5.8.4 Extreme river stages and discharges
Extreme events during floods and droughts should
be documented at both regular gauging stations
and at non-gauged locations.
High-water marks along rivers are useful in delineating flooded areas on maps, in the design of
structures such as highway bridges, and for estimation of flood slopes. These marks, if taken carefully,
may also be used with other data to compute the
peak discharge of the stream by indirect methods
(5.3.5).
Field surveys to measure minimum streamflow
at non-gauged locations provide valuable data at
a very economical cost. These measured
discharges can be correlated with the simultaneous discharges at regular gauging stations to
determine the low-flow characteristics at the
ungauged sites.
2.5.8.5 Video imagery techniques
A video camera installation can provide valuable
information about the conditions at a gauging site.
The extent of ice cover, periods of backwater due to
ice, etc., can be documented by a camera. This technique can also be used for remotely monitoring
potential hazards, for example, risks due to
avalanches.
Recently, video imagery-based approaches have
been used to measure discharge by estimating
surface velocities using particle image velocimetry
methods The video data can be recorded on site, or,
if real-time information is required, readily reported
via some transmission facility.


2.6 MEASUREMENT OF PHYSIOGRAPHIC
CHARACTERISTICS
2.6.1 General
The concepts discussed in this section cover two
quite different physiographic characteristics: the
location of the feature(s) under study, and their
physical response to atmospheric events. By locating these features, it is possible not only to catalogue
them, but also to determine their spatial distribution and the climate zone to which they belong.
The features themselves can be examined in terms
of points, lines, areas or volumes depending on the
relationship between a particular characteristic and
the hydrological regime. For example, streamflow
results from the transformation of climatic events
(rainfall, snowmelt) by the physical complex that
comprises a drainage basin. The basin location
partially determines the climatic characteristics,
which are responsible for meteorological events
that drive the hydrology. However, the basin’s
physical characteristics not only control the hydrological response to the meteorological events, but
some characteristics, for example, orography and
aspect, can also be causal factors in the determination of the basin’s climate.
Physiographic characteristics are now commonly
examined as layers of information within contemporary GIS. The physical response of a watershed to
meteorological events can be analysed using hydrological and hydraulic models as well. The
fundamental procedures presented in this section
form the basis for computer-assisted data assembly
and analysis.


2.6.2 Reference systems and data
frameworks
Physiographic characteristics are but one component of geospatial information; that is, information
pertaining to the character and location of natural
and cultural resources and their relation to human
activity. This information has become so important
that the concepts of national and international
spatial data infrastructure and framework data have
been developed. Spatial data infrastructure can be
considered as the technology, policies, criteria,
standards and people necessary to enable geospatial
data sharing throughout all levels of government, the private and non-profi t sectors, and academia. It
provides a base or structure of practices and relationships among data producers and users that
facilitates data sharing and use. Framework data
can be considered as a set of continuous and fully
integrated geospatial data that provides context
and reference information for the country or region.
In general, this will consist of alignment data such
as geodetic control, data on land features and form
such as physiographic data, and conceptual data
such as government units. A rigorous national data
framework facilitates information exchange and
significantly reduces duplication of effort.


Framework data that will be of interest to hydrological analysis include geodetic control, elevation,
orthoimagery, hydrography, transportation, government units and cadastral information (National
Research Council, 1995).
Geodetic control is defined by using the international system of meridians and parallels divided
into 360 degrees, with the zero meridian passing
through Greenwich. This system is the most widely
used. Its only disadvantage is that a degree in longitude varies from 111.111 km at the Equator to 0 at
the Pole and represents 78.567 km at a latitude of
45° (a degree in latitude always measures 111.111
km). Local systems and other modes of projection
are also in use, for example, the Lambert system.
However, these cannot be recommended in an
international guide. Furthermore, algorithms for
converting geographic coordinates to local reference systems when this may be required are readily
available.


Elevation or altitude is provided in relation to a
given level or reference plane. While local reference
data are sometimes used, until relatively recently
mean sea level was the most commonly used vertical
data. The widespread use of global positioning
system observations led to the adoption of
geocentric vertical (and horizontal) data in
accordance with the world geodetic system, in
preference to those based on mean sea level. The
reference ellipsoid, WGS-84, or a national geocentric
variation is therefore the preferred vertical reference.
The fundamental requirement in any use of a
coordinate system is that the data used must be
specified.


The topography of a river basin may be represented
in two different ways: as a digital elevation model
or as a triangulated irregular network (TIN). The
digital elevation model is a grid of elevation values
that has regular spacing while TIN is a series of
points linked into triangular surfaces that approximate the surface. The spacing of points in TIN are
non-uniform, which allows points to be located on
critical terrain features, roads or river banks. The
accuracy of such digital terrain models depends on
the source of the data, the point density and distribution, and other related data used in their
development. Conventional contour maps may be
prepared from a digital elevation model or TIN.
Orthophotos are images of the landscape from
which features can be referenced to one another.
They are digital images produced by processing
aerial photography to geodetic control elevation
data to remove all sources of distortion. The image
has the properties of scale and accuracy associated
with a map. Such images can be derived using
airborne or satellite sensors.


The basic elements used in estimating physiographic parameters are rarely measured directly by
the hydrologist, who essentially works with global
positioning system data, orthophotos, maps, aerial
photographs and satellite imagery. Therefore, the
accuracy of the evaluation depends upon the accuracy of source materials.


2.6.3 Point measurements
The geometric point is defined here as a unique
location on a line or within an area or volume. A
point may be a physical element, such as the location of a measuring instrument or the outlet of a
basin. It can also be an element of an area (plot of
land) on which a given characteristic or set of characteristics is to be defined or measured. The
physiographic characteristics attributed to a point
may be simple or complex. An example of a simple
characteristic of a point is its elevation, which is
one of its unique identifiers in three-dimensional
space. A more complex characteristic might be a
description of the soil profile that underlies the
point.
Applications of remote-sensing techniques, starting
with aerial photography, has had the effect of
expanding the notion of a point to an area (pixel),
which may measure up to several square kilometres. Within their limits of accuracy, available
techniques may not be able to distinguish between
two points (for example, an instrument’s lack of
resolution), and a pixel might be taken to be a
point.
The horizontal location of a point, that is, its position on the globe, is determined by a selected system
of coordinates (2.6.2), which falls within the scope
of geodesy and topography. A universal system has
been invented to make the coding of a point in a catalogue explicit by indicating its geographical
position. This is the GEOREP squaring system
(UNESCO, 1974) for spatial representation of linear
features. Other systems may locate points by their
linear distances along a stream from a given origin,
for example, mouth or confluence.

The physiographic description of a point covers its
geometric properties (form, relief, slope, etc.) and
its permanent physical properties (permeability,
nature of rocks, soil structure, land-use type, etc.).
The former are limited to the local slope, while the
latter comprise a whole range of possible physical
properties, expressed in scalar form for a point on a
horizontal surface or in vectorial form for a profile,
for example, geological core.

2.6.4 Linear measurements
Any physiographic element is linear if it can be
represented by a line on a map or in space. In hydrology, three types of linear elements are common:
(a) Boundaries;
(b) Isopleths of a permanent feature, for example,
contours;
(c) Thalwegs.
The first two types are linked to areal aspects, which
will be examined later.
The thalweg is itself to be considered not only as
represented in horizontal projection and longitudinal profile, but also by the way in which it combines
with other thalwegs to form a drainage network,
which has its own physiographic characteristics.
Some drainage network characteristics are linear,
for example, the bifurcation ratio, while others are
areal in nature, such as the drainage density.
2.6.4.1 The stream
A stream in horizontal projection may be represented, if the scale of the diagram is suitable, by two
lines representing its banks. From these two lines,
an axis can be drawn equidistant to the two banks.
The axis may also be defined as the line joining the
lowest points on successive cross-sections. In fact,
these elements, the visible banks and the lowest
points, are not always very clear, and the map scale
does not always permit the banks to be featured
properly. Mapping, thus, is reduced to representing
a stream by a line.


Lengths along a river are measured by following
this line and by using a curvometer. The accuracy of
the determination depends on the map’s scale and
quality, as well as on the curvometer’s error, which
should not exceed six per cent for a distance on the
map of 10 cm or 4 per cent for 100 cm and 2 per
cent beyond. Many hydrological features can be
derived directly from the orthoimagery or digital
terrain data with the aid of GIS (2.6.7).
The axis of a stream is rarely straight. When it
comprises quasi-periodic bends, each half-period
is called a meander. The properties and dimensions of meanders have been thoroughly studied
by geographers and specialists in river
hydraulics.


2.6.4.2 The drainage network
In a basin, streams are organized to form a drainage
network. In a network, all streams are not the same
size, and several systems have been proposed for
classifying them. Several stream classification
systems are in use in various countries and current
GIS provide for automatic stream classification
according to schemes devised by Horton, Schumm,
Stahler, Shreve and others. The best known schemes
is Horton’s, in which any elementary stream is said
to be of order 1, any stream with a tributary of order
1 is said to be of order 2, and any stream with a
tributary of order x is said to be of order x + 1. At a
confluence, any doubt is removed by giving the
higher order to the longest of the tributaries forming it (Figure I.2.15) (Dubreuil, 1966). This
introduces some inaccuracy that was avoided by
Schumm by systematically giving order x to the
reaches formed by two tributaries of order x – 1
(Figure I.2.16). The main source of error in such
evaluations is to be found in the mapping of the
streams, where the definition of the smallest streams
is often rather subjective.


Of the linear characteristics of the drainage network
that are measurable on a map, the confluence ratio
Rc and the length ratio Rl are based on Horton’s laws
and have been verified for Horton’s classification.
Given that Nx is the number of streams of order x,
and lmx=∑l x/Nx is the mean length of the streams of
order x, these laws are expressed by the following
relationships:

Figure I.2.15. Horton’s classifi cation
Figure I.2.16. Schumm’s classification

2.6.4.3 Stream profile
The stream profile is the variation in elevation of
the points of the stream thalweg as a function of
their distance from the origin, which is generally
taken as the confluence of the stream with a larger
stream or as its mouth. On such a profile, a certain
number of topographical features are to be found,
such as high points (thresholds), hollows between
two thresholds (pools), rapids, waterfalls and
changes of slope that frequently mark the boundary between two reaches with different geologic
controls (Figure I.2.17).


The average slope of a whole stream is the difference in elevation between its highest point and its
confluence or mouth divided by its total length.
This notion is simple, but not very useful. On the
other hand, knowledge of the slopes of the successive stream reaches is essential for most runoff and
hydraulic models.
The profiles of the main stream and of various tributaries in the same basin can be represented on the
same diagram. Figure I.2.18 shows examples of
stream profi les of the Niger river at Koulikoro and
of its main tributaries and sub-tributaries. Such a
diagram gives a synthesized view of the variation in
slope of the drainage network’s elements.


2.6.4.4 Cross-section
The profile of the valley taken perpendicular to a
stream’s axis is called a cross-section, and a series of
these is valuable information for the development
of streamfl ow models. Cross-sections are used in
several types of calculations, and the way in which
they are established may depend on the use to
which they will be put.
An important particular case is the calculation of
fl ow for a discharge measurement, in which elevation is expressed as a depth and is obtained by
sounding (5.3). Cross-sections are usually obtained
by making normal topographical measurements
during the lowest flows.
2.6.4.5 Physical characteristics
The type of material in the stream bed (particularly
its cohesiveness), the type and amount of vegetation in and along the stream, and the roughness of
the bed, which depends on the longitudinal and
transverse distributions of the former, comprise the
primary physical characteristics of a stream.

Figure I.2.17. Stream profile
Figure I.2.18. Profile of the Niger river and its tributaries


2.6.5 Area measurements
2.6.5.1 The basin
The basin is defined as the area that receives precipitation and, after hydrological processes resulting
in losses and delays, leads it to an outflow point.
The watershed boundary, the basin’s perimeter, is
such that any precipitation falling within it is
directed towards the outflow, whereas any precipitation falling outside drains to a different basin and
outfl ow. In some cases, it may not be easy to determine the basin boundary, for example, when the
head of the main stream is formed in a very fl at bottomed valley or a marshland. The watershed is
usually defined by using contour maps or aerial
photographs.
The basin perimeter is measured in a GIS (2.6.7)
or with a curvometer. The measured perimeter is
a function of the scale and accuracy of the maps
or photographs, the quality of the curvometer,
and the care taken in its use (Figure I.2.19). The
ultimate use that will be made of the measurement should determine the accuracy to which it
is measured.
The basin area is determined in a GIS or measured
by planimetry by following the boundaries established as described above.

Figure I.2.19. Real and measured perimeter


The basin’s shape is characterized by comparing its
perimeter with that of a circle having the same
area. If A is the basin area and P its perimeter, both
measured according to the above rules and
expressed in compatible units, then the ratio of the two perimeters is called the Gravelius coefficient of
compactness, which is given by:

C = 0.282 P A1/2

The notion of an equivalent rectangle is also linked
to the basin’s shape, and permits the definition of a
particular slope index. The equivalent rectangle has
the same area and the same Gravelius coefficient as
the basin. The length of this rectangle is:

The drainage density is defined as the total length of
streams of all orders contained in the basin’s unit area:

Dd = (∑Lx)/A

where Lx is the total length of the streams of order
x. In common practice, the lengths are expressed in
kilometres and the areas in square kilometres.

Figure I.2.20. Relief and drainage network
(Courtesy ARPA-Piemonte)


The basin relief, shown on maps by contours, can
be described by the hypsometric distribution or the
hypsometric curve. Figure I.2.20 shows a representation of the relief and drainage network. The elevation ranges are shown by different marking.

The hypsometric distribution gives the percentage
(or fraction) of the basin’s total area that is included
in each of a number of elevation intervals. The
hypsometric curve shows, on the ordinate, the
percentage of the drainage area that is higher than
or equal to the elevation that is indicated by the
corresponding abscissa (Figure I.2.21). In practice,
the cumulative distribution of area is obtained in a
GIS or by planimetric calculation of successive areas
between contours of elevation beginning with the
basin’s lowest point.


It is possible to calculate the basin’s mean elevation
by dividing the area under the hypsometric curve
by the length of the ordinate corresponding to the
whole basin.


The basin slope can be represented by several indices. The oldest, and perhaps still the most widely used, is the basin’s mean slope Sm. It is determined from the basin contours by the formula:

Figure I.2.21. Hypsometric curves
(Courtesy ARPA-Piemonte)

Where z is the contour interval, ∑l is the total length
of all contours within the basin, and A is the basin’s
area. The difficulty and main source of error in estimating this characteristic lie in the measurement of
∑l. The contours are almost always very tortuous
and their real length is not really characteristic of
the role they play in calculating the index. Therefore,
it may be necessary to smooth the irregularities
keeping in mind the final results may be somewhat
inconsistent and variable.


A mean slope can also be estimated by taking the
basin’s total difference in elevation and by dividing it by one of its characteristic dimensions.
However, the distribution of slopes in the basin is
neglected by this approach. One way of avoiding
this is to derive the slope index from the hypsometric curve, which is a synthesis of the relief
delineated by the contours, and to weigh the
areal elements corresponding to the various elevation intervals by a non-linear function of the
mean slope in each interval. Roche’s slope index,
also called the index of runoff susceptibility,
meets these conditions. The notion of the equivalent rectangle (equation 2.14) is applied to each
contour to transform geometrically the contours
into parallel straight lines on the rectangle representing the entire basin (Figure I.2.22).

If ai and ai–1 are the elevations of two successive contours
and xi is the distance separating them on the
equivalent rectangle, the mean slope between
these two contours is taken to be equal to
(ai– ai–1) / xi, and the slope index is written by
designating as ñithe fraction of the basin’s total
area included between ai and ai–1:

The Roche slope index is as follows:

When basins have a very low slope, for example, in
the interior plains of North America, there may be
closed sub-basins having no outlet to the main
stream or significant portions of the basin that
contribute to streamflow very infrequently.

Under these circumstances the concept of an effective
drainage area may be used. This is customarily
defined as the area that would contribute to streamflow in a median year. Establishing the effective drainage area for a basin may require significant cartographic and hydrological analysis.


A basin’s physical characteristics are essentially the
soil types, the natural plant cover or artificial cover
(crops), the land cover (for example, lakes, swamps,
or glaciers), and the type of land use (for example,
rural or urbanized areas, lakes, or swamps). They
may also be expressed in terms of the basin’s reaction to precipitation, this is, classes of permeability.

These physical characteristics may be assembled as
layers within a GIS.


The quantification of these characteristics requires
definition of criteria and procedures for delineating the areas meeting these criteria. It then remains
only to measure each of these areas and to express
each as a percentage (or fraction). The tools for
determining such distributions include GIS, normal
and/or specialized cartography, aerial photography
and remote-sensing with relatively fine resolution
(pixels not to exceed some hundreds of square
metres).


2.6.5.2 The grid
The formation of physiographical data banks,
especially for the development of rainfall-runoff
models with spatial discretization, leads to the
division of the basin area based on systematic
squaring or griding. Depending on the objective,
the grid size may be larger or smaller, and may be
measured in kilometres (1 or 5 km2) or based on
the international geographical system (1’ or 1°
grid). GIS (2.6.7) have made interchanging
between gridded and ungridded data a simple
task once the initial databases have been
assembled.
2.6.6 Volumetric measurements
Volumetric measurements pertain primarily to
the defi nition of water and sediment storage.
Evaluation of groundwater storage is covered by
hydrogeology. It therefore will not be discussed
here, nor will the estimation of sediment deposited on the soil surface. Surface storages are
generally either the volumes of existing lakes or
reservoirs, for which bathymetric methods are
used, or the volumes of reservoirs that are being
designed, for which topographical methods are
used.
2.6.6.1 Bathymetric methods
Ordinary maps rarely give bathymetric data on
lakes and reservoirs. The volume of an existing
reservoir, therefore, has to be measured by making
special bathymetric readings. Usually, this is done
from a boat by using normal methods for sounding and for positioning the boat. The depths
should be referenced to a fi xed datum and a stage
gauge or a limnigraph so that variations in stage
can be monitored.
Depth measurements can be used to plot
isobaths, and the reservoir’s volume above a
reference plane can be calculated through
double integration (generally graphical)
of the isobath network. One application of this
method is sedimentation monitoring in a
reservoir.
2.6.6.2 Topographical methods
Once the site of a dam has been fi xed, the calculation of the reservoir’s effi ciency and management
requires knowledge of the curve of volume
impounded as a function of the reservoir’s
stage (stage-volume curve). To determine this relationship, ground-surface-elevation contours are
needed throughout the area to be occupied by the
future impoundment. This requires maps or topographical plans of the area on scales of between
1/1 000 and 1/5 000. If these are not available, maps
on a scale of 1/50 000 can be used for preliminary
design, but a topographical survey on an appropriate scale will be needed subsequently.


By using the contour map, planimetric
measurements are made, in a GIS or manually, of
the areas contained within the contours with the
hypothetical reservoir in place. A plot of these
areas versus their related elevations is known as a
stage-area curve. The stage-volume curve is
computed from the stage-area curve by graphical
integration.
2.6.7 Geographical Information Systems
GIS are now ubiquitous in the fi elds of operational
hydrology and water resources assessment. Many
aspects of data collection and interpretation can be
facilitated by means of GIS.
In network planning and design, the ability to map
quickly and display surface water and related
stations enables a more effective integration to take
place. Network maps, showing basins or stations
selected according to record quality, watershed, or
operational characteristics, can be used for both
short-term and long-term planning. The essential
features of complex networks can be made very
clear.
GIS techniques are being incorporated in hydrological models for the purpose of extracting and
formatting distributed watershed data. Used in
conjunction with digital elevation models or TINs
(2.6.2), complete physiographic and hydrological
depiction of basins can be readily accomplished.
Runoff mapping and interpolation is being carried
out using GIS routines in many countries. The effi –
ciency of handling large volumes of data means
that more comprehensive and detailed maps,
isolines and themes can be prepared. This represents a significant improvement to water
resources-assessment technology, as map preparation is often time-consuming and expensive.

Figure I.2.22. Equivalent rectangles

The interpretation of real-time data can also be
facilitated through GIS. The thematic mapping of
stations reporting over threshold amounts or digital
indications of rainfall would obviously be very
useful to both operational hydrology and forecasting agencies.
GIS systems are now available for standard computers in practical, low-cost formats. The main cost
factor now resides in the areas of database compilation, and training and updating of technical
staff.
2.6.8 Emerging technologies
The subsequent chapters of this Volume of the
Guide deal with proven technologies that are
commonly used in many parts of the world.
However, as indicated above, new technologies are
continuously evolving. This section provides some
insight into several of these so that Hydrological
Services may be kept aware of their possibilities.
2.6.8.1 Remote-sensing
In the field of hydrological measurements, two
kinds of remote-sensing techniques are commonly
used: active (by emission of an artificial radiation
beam toward the target and analysis of the target
response), or passive (by analysis of the natural
radiation of an object).
In active methods, radiation may be high-frequency
electromagnetic (radar) or acoustic (ultrasonic
devices). The apparatus may be installed on the
ground (radar, ultrasonic), on airplanes, or on satellites (radar). Active remote-sensing is usually done
on an areal basis, but may also be used for pointoriented measurements (ultrasonic).
In passive methods, the radiation is electromagnetic (from infra-red (IR) to violet, and rarely
ultraviolet). Most current applications are made by
means of a multi-spectral scanner, which may be
airborne, but is more frequently carried on a satellite. Passive sensing is always areal.
Radars are now used for quantitative precipitation
estimates over a given area. Snow-water equivalent
can be determined by measuring the natural gamma
radiation from potassium, uranium and thorium
radioisotopes in the upper 20 cm of soil under bare
ground conditions and with the snowpack.
Observations are made from a low-flying aircraft.
Data are collected on a swath about 300 m wide and
15 km long. Results will be affected by ice lenses or
liquid water in the snowpack, ground ice or
standing water (Carroll, 2001). Microwave sensors,
both airborne and satellite, have been used as well
to monitor snowpack properties. RadarSat active
radar has also been used to map the areal extent of
wet snow.

Airborne optical devices (Lidar) are now used to
determine topography more rapidly and, often,
more accurately and at lower cost than conventional aerial photography. The resulting digital
elevation model has applications in hydraulic and
hydrological modelling and in determining glacier
mass balance. Satellite Lidar altimetry has been
used to obtain very good topography for military
purposes and in research applications, but has not
yet been commercialized. In the absence of national
topographic data, the low-resolution global digital
elevation model GTOPO30 with a horizontal grid
spacing of 30 arc seconds (roughly 1 km) may be
considered. The vertical accuracy of the data is
about 30 m. This digital elevation model is also
linked to the HYDRO1k package which provides a
suite of six raster and two vector data sets. These
data sets cover many of the common derivative
products used in hydrological analysis. The raster
data sets are a hydrologically correct digital elevation model, derived flow directions, flow
accumulations, slope, aspect and a compound topographic (wetness) index. The derived streamlines
and basins are distributed as vector data sets.


A further existing topographic data option is the
3 arc-second (90 m) digital elevation model
produced by the Shuttle Radar Topography Mission.
The data for most of the coverage area have been
processed to level 1, which provides for an absolute
horizontal accuracy of 50 m and a vertical accuracy
of 30 m. The level 2 digital elevation model,
currently available only for the United States, has a
horizontal accuracy of 30 m and vertical accuracy
of 18 m.


Other uses of remote-sensing in hydrology include
sensing of near-surface soil moisture using airborne
natural gamma or satellite passive microwave techniques and measurement of land surface temperature
as a precursor to determining evapotranspiration.
Leaf area index measurements use may also lead to
remote-sensing of evapotranspiration. Remotesensing of water quality also offers considerable
promise as new satellites and sensors are developed.

Water bodies that are affected by suspended sediment, algae or plant growth, dissolved organic
matter, or thermal plumes undergo changes in spectral or thermal properties that may be detected by
airborne or satellite sensors (UNEP/WHO, 1996).
Some use has been made in the measurement of water body areas and the extent of flood inundation using RadarSat active radar. Aside from the requirement to calibrate airborne or satellite sensors,
there is also a need to ground-truth the remotely
sensed data to ensure that remotely sensed values
represent in situ values.
2.6.8.2 Hydroacoustic methods
Hydroacoustic methods hold considerable promise
for hydrological data acquisition. Acoustic signals
may be used to identify the interface between two
dissimilar media or to explore the characteristics of
a single medium. For example, echo sounders are
used to define the streambed in hydrographic
surveys or to sense the distance to the water surface
when mounted in or above a stream. Results can be
very satisfactory provided careful attention is paid
to calibration of the instrument. Acoustic current
meters that determine water velocity by measuring
the Doppler shift of acoustic energy reflected from
water-borne particles have been used for a number
of years.

The 1990s saw the development of the Acoustic
Doppler Current Profiler (ADCP), an instrument
that uses acoustic energy to determine streamflow ow
from a moving boat. The instrument consists of
four orthogonal ultrasonic transducers fixed to a
moving boat. As the boat traverses a river the
instrument measures the frequency shift of the
reflected signals and uses trigonometry to produce
velocity vectors in uniformly spaced volumes
known as depth cells. The velocity of the boat is
removed in computer processing and, with the
channel geometry also defi ned by the instrument,
the streamflow along a river transect can be
calculated. This technique has been used successfully
to measure relatively large streams. More recently,
efforts have been directed to the measurement of
smaller streams (under 2 m depth) using hand-held
or in situ instruments.


Acoustic devices have also been developed to examine lake dynamics or to determine the density and
material characteristics of bottom and sub-bottom
sediments. Ultrasonic flowmeters are reviewed in
Chapter 5.
2.6.8.3 Risk reduction for personnel
There are inherent dangers to personnel involved
in acquiring hydrological data under diffi cult conditions. These dangers are perhaps best exemplified
by the challenge of measuring streamflow under
fl ood conditions. High velocities, debris or ice may
threaten the life of persons attempting to make the
measurement. Efforts are therefore underway to
automate the measurement process through use of
robotics and other procedures. One early approach
to improved safety was the development of streamgauging cableways that could be operated from the
river bank. Another was the moving boat method,
which reduces the time required for a discharge
measurement, but still requires exposing personnel
to the hazard.

One current concept calls for an automated,
unmanned boat equipped with an Acoustic Doppler
Current Profi ler the position of which is monitored
by use of the global positioning system.
Measurements can therefore be made under high
hazard conditions with minimum exposure of
personnel to the hazard. Another approach uses a
hand-held radar to measure surface velocities and,
where channels are unstable, ground-penetrating
radar to defi ne the channel cross-section. The radar
device produces an accurate surface velocity, which
must then be related to mean velocity, while the
ground-penetrating radar moving along a bridge or
cableway produces an accurate cross-section.
Other risk reduction efforts include the decommissioning of water-level sensors based on mercury
manometers and the increased use of satellite telephones as a means of maintaining contact with
fi eld parties in remote areas.
2.6.9 Staff training
Whatever the level of technical sophistication of a
data-collection authority, the quality of its staff will
always remain its most valuable resource. Careful
recruitment, training and management is the key
to attaining and maintaining the appropriate
personnel.
WMO has published a set of Guidelines for the
Education and Training of Personnel in Meteorology
and Operational Hydrology (WMO-No. 258). UNESCO
has published a document on Curricula and Syllabi
in Hydrology (UNESCO, 1983). With respect to data
collection and processing, employee education,
although costly and time-consuming, can be a
sound investment that results in greater productivity and effectiveness. A carefully structured training
programme is essential for all personnel engaged in
data collection because they are in a strong position
to influence the standard of the final data. Formal
training should aim at providing both a general
course in first principles, plus training modules to
teach in-house procedures. All material should be
relevant and current. The Canadian hydrometric
technician career development programme (HOMS component Y00.0.10) provides one national
example (WMO, 2000). Volume II, Chapter 2,
provides additional information on different aspects
of training in hydrology.
Where processing is not carried out by the data
collector, it is important that data processors be
trained in data-collection techniques to ensure that
data are processed according to the intent of the
collector. It is a good practice to give processing
staff periodic fi eld experience to build a physical
association with the data and their origins. Such
knowledge on the part of the processor can allow
interim interpretations of incorrectly presented
data, pending confirmation from the collector. It is
essential to establish the principle that the person
collecting the data has the primary responsibility
for its quality. One method of honouring this principle is to involve the collector in the processing as
much as possible, and to ensure that feedback is
obtained by returning the published data to the
collector for assessment. At the processing stage,
staff should recognize that they also have a responsibility to maintain the quality and integrity of the
data.
Data processing is often routine in nature and well
suited to the application of automation and technology. For this reason, it is important that special
attention be given to the care of human resources,
and that the system be structured to foster interest,
involvement, professionalism and a sense of
achievement. Data-processing staff should be given
the opportunity to contribute ideas that may
increase the effectiveness of the processing system.
Staff safety is also an integral component of any
profession, and the duties undertaken by data
collectors and processors require the establishment
of safety standards. These are primarily discussed in
Chapter 8. However, the possibility of repetitive
strain injury in data-processing staff can often be
caused by routine and the repetitive nature of some
aspects of their jobs. This problem should be
addressed from both a staff safety and a management point of view.

References and further reading
Azar, J., D. Sellars and D. Schroeter, 2003: Water
Quantity Monitoring in British Columbia:
A Business Review of the BC Hydrometric
Programs. British Columbia Ministry of
Sustainable Resource Management, Victoria,
BC (http://www.geoscientific.com/technical/
tech_references_pdf_files/Water%20Quantity%20M
onitoring%20in%20BC.pdf).
Carroll, T., 2001: Airborne Gamma Radiation Snow Survey
Program: A User’s Guide. Version 5.0. National
Weather Service, Chanhassen, Minnesota (http://
www.nohrsc.noaa.gov/technology/pdf/tom_
gamma50.pdf).
Carson, M.A., 1987: An Assessment of Problems Relating
to the Source, Transfer and Fate of Sediment along
the Mackenzie River, NWT. Internal Report, Water
Resources Branch, Environment Canada.
Church, M.A., R. Kellerhals and T.J. Day, 1989: Regional
clastic sediment yield in British Columbia. Canadian
Journal of Earth Sciences, Volume 26,
No. 1, pp. 31–45 (http://cgrg.geog.uvic.ca/abstracts/
ChurchRegionalThe1989.html).
CNS Scientific and Engineering Services, 1991: The
Benefit-cost of Hydrometric Data: River Flow Gauging.
Report No. FR/D0004, Foundation for Water
Research, Marlow (http://www.fwr.org/urbanpol/
frd0004.htm).
Dubreuil P., 1966: Les caractères physiques et
morphologiques des bassins versants: leur détermination
avec une précision acceptable. Cahiers ORSTOM, Série
Hydrologie, No. 5.
Environment Canada, 1983: Sampling for Water Quality.
Water Quality Branch, Inland Waters Directorate.
Environment Canada, Ottawa.
Harvey, D.K.D., P.J. Pilon and T.R. Yuzyk, 1999:
Canada’s Reference Hydrometric Basin Network
(RHBN). Proceedings of the Fifty-first Annual
Conference of the Canadian Water Resources
Association, Halifax.
International Organization for Standardization, 1993:
ISO Standards Handbook: Quantities and Units. Third
edition, Geneva (http://www.iso.org/iso/en/prodsservices/otherpubs/pdf/quantity1993-en.pdf).
International Organization for Standardization,
2005: Measurement of Fluid Flow: Procedures for
the Evaluation of Uncertainties. Second edition,
ISO 5168, Geneva. (http://www.iso.org/iso/en/
CatalogueDetailPage.CatalogueDetail?CSNUMBER=
32199&ICS1=17&ICS2=120&ICS3=10&showrevisio
n=y).
International Organization for Standardization, Technical
Committee 147 List of Standards on water quality
(http://www.iso.org/iso/en/stdsdevelopment/
tc/tclist/TechnicalCommitteeStandardsListPage.
TechnicalCommitteeStandardsList?COMMID=3666
&INCLUDESC=YES).
International Union for the Conservation of Nature and
Natural Resources, United Nations Environment
Programme and World Wildlife Fund, 1991: Caring
for the Earth: A Strategy for Sustainable Living. Gland
(http://gcmd.nasa.gov/records/GCMD_IUCN_
CARING.html)

Moss, M.E. and G.D. Tasker, 1991: An intercomparison of
hydrological network design technologies. Hydrological
Science Journal, Volume 36, No. 3, pp. 209–221.
National Research Council, 1995: A Data Foundation for
the National Spatial Data Infrastructure. Commission
on Geosciences, Environment and Resources,
National Academy of Sciences, Washington DC
(http://darwin.nap.edu/books/NX005078/html/
index.html).
Pearson, C.P., 1998: Changes to New Zealand’s National
Hydrometric Network in the 1990s. Journal of Hydrology,
Volume 37, No. 1, pp. 1–17.
United Nations, 1992a: International Conference on Water
and the Environment: Development Issues for the
Twenty-first Century, 26–31 January 1992, Dublin
(http://www.wmo.int/pages/prog/hwrp/documents/
english/icwedece.html).
United Nations, 1992b: Report of the United Nations
Conference on Environment and Development. Rio de
Janeiro, 3–14 June 1992, United Nations publication, Sales No. E.93.I.8, Three Volumes, United
Nations, New York (http://www.centrodirittiumani.
unipd.it/a_temi/conferenze/rio/Dicrio1992.pdf).
United Nations, 1997: Comprehensive Assessment of the
Freshwater Resources of the World. New York.
United Nations, 2002: Report of the World Summit on
Sustainable Development. Johannesburg, South Africa,
26 August to 4 September 2002 (http://www.unctad.
org/en/docs/aconf199d20&c1_en.pdf).
United Nations, 2004: Guidelines for Reducing Flood
Losses. United Nations, New York (http://www.
unisdr.org/eng/library/isdr-publication/floodguidelines/Guidelines-for-reducing-floodslosses.pdf).
United Nations Educational, Scientific and Cultural
Organization, 1974: The GEOREP grid station identification system. In: Discharges of Selected Rivers of the
World, Volume III, Part II, 1969–1972.
United Nations Educational, Scientific and Cultural
Organization, 1983: Curricula and Syllabi in Hydrology
(S. Chandra, L.J. Mostertman, J.E. Nash, J. Nemec
and T. Peczely). Technical Papers in Hydrology,
No. 22, Paris.
United Nations Educational, Scientific and Cultural
Organization and World Meteorological
Organization, 1988: Water-resource Assessment
Activities: Handbook for National Evaluation. (http://
unesdoc.unesco.org/images/0015/001584/158461eo.
pdf).
United Nations Environment Programme Global
Environment Monitoring System (GEMS)/Water
Programme, 2005: Global Environment Monitoring
System (GEMS)/Water Operational Guide. Fourth
edition, Inland Waters Directorate, Burlington,
Ontario (http://www.gemswater.org/publications/
operational_guide.html).
United Nations Environment Programme and World
Health Organization, 1996: Water Quality Monitoring:
A Practical Guide to the Design and Implementation of
Freshwater Quality Studies and Monitoring Programmes.
Jamie Bartram and Richard Balance (eds), E&FN
Spon, London (http://www.who.int/water_sanitation_health/resourcesquality/waterqualmonitor.pdf).
World Meteorological Organization, 1969: Hydrological
Network Design: Needs, Problems and Approaches
(J.C. Rodda). World Meteorological Organization
and International Hydrological Decade Projects
Report No. 12, Geneva.
World Meteorological Organization, 1972: Casebook on
Hydrological Network Design Practice. WMO-No. 324,
Geneva.
World Meteorological Organization, 1976: Hydrological
network design and information transfer. Proceedings
of the International Seminar, 19–23 August 1974,
Newcastle-upon-Tyne, Operational Hydrology
Report No. 8, WMO-No. 433, Geneva.
World Meteorological Organization, 1980: Manual
on Stream Gauging. Volumes I and II, Operational
Hydrology Report No. 13, WMO-No. 519, Geneva.
World Meteorological Organization, 1981: Hydrological
Data Transmission (A.F. Flanders). Operational
Hydrology Report No. 14, WMO-No. 559, Geneva.
World Meteorological Organization, 1982: Concepts and
Techniques in Hydrological Network Design
(M.E. Moss). Operational Hydrology Report No. 19,
WMO No. 580, Geneva.
World Meteorological Organization, 1987: Hydrological
Information Referral Service: INFOHYDRO Manual.
Operational Hydrology Report No. 28, WMONo. 683, Geneva.
World Meteorological Organization, 1988a: Manual
on Water Quality Monitoring: Planning and
Implementation of Sampling and Field Testing.
Operational Hydrology Report No. 27, WMONo. 680, Geneva.
World Meteorological Organization, 1988b: Technical
Regulations. Volume III, Hydrology, WMO-No. 49,
Geneva.
World Meteorological Organization, 1990a: Cost–benefit
Assessment Techniques and User Requirements for
Hydrological Data. Operational Hydrology Report
No. 32, WMO-No. 717, Geneva.
World Meteorological Organization, 1990b: Economic
and Social Benefits of Meteorological and Hydrological
Services. Proceedings of the Technical Conference,
26–30 March 1990, Geneva, WMO-No. 733, Geneva.
World Meteorological Organization, 1992: Proceedings
of the International Workshop on Network Design
Practices. 11–15 November 1991, Koblenz,
Technical Report No. 50, WMO/TD-No. 671,
Geneva.
World Meteorological Organization, 1994: An Overview
of Selected Techniques for Analysing Surface-water Data
Networks (W.O. Thomas). Operational Hydrology
Report No. 41, WMO-No. 806, Geneva.

World Meteorological Organization, 1995, 2001, 1998:
Manual on Codes. Volumes I and II, WMO-No. 306,
Geneva.
World Meteorological Organization, 1996: Guide to
Meteorological Instruments and Methods of Observation.
Sixth edition, WMO-No. 8, Geneva.
World Meteorological Organization, 1998a: Current
Operational Applications of Remote Sensing in Hydrology
(A. Rango and A.J. Shalaby). Operational Hydrology
Report No. 43, WMO-No. 884, Geneva.
World Meteorological Organization, 1998b: WHYCOS:
World Hydrological Cycle Observing System.
WMO-No. 876, Geneva.
World Meteorological Organization, 2000: Hydrological
Operational Multipurpose System (HOMS) Reference
Manual. Second edition, Geneva.
World Meteorological Organization, 2002, 2003:
Guidelines for the Education and Training of Personnel
in Meteorology and Operational Hydrology. Volumes I
and II, Fourth edition, WMO-No. 258, Geneva.
World Meteorological Organization, 2003: Manual on the
Global Observing System. Volume I, WMO-No. 544,
Geneva.
World Meteorological Organization, 2004: Water
resources as a challenge of the twenty-first century
(M. Abu-Zeid and. I.A. Shiklomanov). WMONo. 959, Geneva.
World Meteorological Organization and United Nations
Educational, Scientific and Cultural Organization,
1991: Progress in the Implementation of the Mar del
Plata Action Plan and a Strategy for the 1990s. Report
on Water Resources Assessment, New York.