Contents - Previous - Next

4.3 Data processing, formats for data recording

4.3.1 Kinds of data to be processed

There are several sources of information from the measurement programme:

• continuous precipitation, water stage height, and timing of automatic sampling, from the data-loggers. These are typically in the form of 5-minute step or variable-time step computer files that allow a full computerized processing, with excellent advantages compared with the traditional hand-graphic processing.

• instantaneous water discharge measurements made by an operator with current-meter or chemical methods. These data are identified by date and time as well as water stage.

• instantaneous sediment concentrations during the rising limb of the hydrograph, obtained through the analysis of samples taken by the stage samplers. These data are identified by the corresponding water stages in the rising limb of the hydrograph.

• instantaneous sediment concentrations during flood events, obtained through the analysis of samples taken by a programmable automatic sampler or manually by an operator. These data are identified by the date and time recorded by the data-logger or the operator.

4.3.2 Steps for data processing

There are two kinds of information during the data processing: actual data, obtained from the four sources listed in the former paragraph, and derived data, obtained through some computation with a stage-discharge relationship or a water-sediment discharge rating formula. Original actual data are to be protected against loss or deterioration, and derived data are to be considered as open to improvement, if better relationships are obtained during the programme.

The steps needed to process data are the following ones:

a) Establishment of 'stage-discharge' relationships or rating formulae for the gauging stations. The recommendations of the former chapter are that these relationships should be provided by the designer of the runoff control, if these are artificial controls, or by the water authorities, if these are pre-existing runoff stations. If these recommendations cannot be fulfilled and new control stations with unknown rating formula are set up, the best solution is to take some usual formula (critical flow, Manning-Strickler...) with provisional parameters, and to improve it during the measurement programme through performing flow measurements with a current meter for a range of discharges as wide as possible. Nevertheless, it is worth stating that stage-discharge relationships do not provide usually the main source of errors for estimating sediment yield.

b) Establishment of chronological data files containing raw data from the data loggers and derived data computed through calibration or rating formulae. This is recommended to be made by using some computer spreadsheet software, like EXCEL or LOTUS 1-2-3. These files should be set up shortly after data retrieval in order to check any malfuntioning of instruments, but early files can be updated or improved if necessary with sediment rating formulae obtained later during the Programme. Time integration of water and sediment discharges are to be made in the same spreadsheets at the data-logger time step.

c) Establishment of sediment discharge versus water discharge relationships. It is first necessary to determine the sediment concentration of the samples, and to attribute a date, time, water stage and water discharge to every sediment concentration. The analysis of the sediment-discharge relationship, that may be made with the help of a computer spreadsheet, is crucial for understanding the erosion and transport processes as well as for obtaining an adequate sediment rating procedure that will allow a reliable assessment of sediment yield. Using a unique sediment rating formula for all the water discharge values obtained at a station should be avoided within the Programme, because errors could be very important. The best solution would be to use two different sediment rating formulae for the rising and recession limbs respectively of every significant event, obtained from sediment data collected for the same event. Simpler solutions could be to use rating formulae adapted to different seasons and hydrograph limbs.

d) Production of temporal graphs of instantaneous precipitation, water discharge, suspended discharge and dissolved discharge for the whole period of acquisition, as well as more detailed graphs for the main relevant events. Production of tables with the time-integrated volumes of precipitation, runoff, suspended sediment yield and dissolved sediment yield for the main events, total period of acquisition and annual area-averaged rates.

Figure II-13: Example of spreadsheet for handling sampling data. Sample order should be kept in chronological order, in order to allow the study of the likely hysteresis behaviour. A column can be added for remarks.

4.3.3 Sample data spreadsheets

There are two different kinds of data to be handled with the help of spreadsheets: sampling data, obtained by hand or automatic sampling, and chronological data, obtained by the data-loggers.

Sampling data: Figure II-13 displays an example of the recommended procedure to handle sample data on a computer spreadsheet. This spreadsheet can be also used to analyze and to build the relationships stage - discharge, electrical conductivity- dissolved concentrations, and water discharge sediment discharge.

Nevertheless, the usual spreadsheets do not allow the study of a wide range of curve-fitting techniques, and the usual way to fit a potential function is the linear regression of the logarithmic values. In these conditions, the best solution is to perform the following steps, for every separate event, if possible:

a) Make two columns with the logarithms of water discharges and solid discharges (product between water discharge and sample solids concentration) corresponding to all the available samples.

b) Build the XY graphs for actual water and sediment discharges as well as for the logarithmic values, keeping the chronological order of sampling. Study the hysteresis loops and the relative significance of every sample in the total transport (actual values). If hysteresis phenomena are significant, make the following steps for every different part of the hydrograph (rising and recession limbs usually).

c) Compute both linear regressions with actual and logarithmic values, build three new columns with the predicted values from the two regressions and with the antilogarithms of the values predicted by the logarithmic regression, display them on the respective graphs.

d) Compute and cumulate the differences between measured and predicted sediment discharges, using the anti-logarithms of the log-log regression. If the difference is considered too big, select a smaller number of samples (usually the problem is that there are too many samples for very low discharges), and repeat the former step (c).

e) Select the set of the more adequate regressions for predicting sediment discharges, the more intricate solution being a different equation for every part of the hydrograph of every event, and the more simpler one, a simple regression for all the observed data. Establish a table with the differences observed with the several methods (see Table II-1).

Chronological data files cannot easily be handled with a spreadsheet if all the month is represented by a 5-minute step (a constant step data logger is used). There are two solutions to this problem: rows with repeated values of stage height without rainfall or sampling data can be deleted, or 15-days periods of full 5-minutes step are used. Temporal integration of water and sediment discharges must be given in the same spreadsheet, by multiplying instantaneous discharges by the time interval between readings (rows).

Figure II-14: Example of spreadsheet for handling chronological data logger information. A supplementary column can be added for remarks.

Figure II-14 displays an example of the spreadsheet that can be used for handling chronological data. Sediment discharge and transport columns can be made for alternative sediment rating equations. A row at the end of the sheet can be added for total cumulated volumes of water and sediment.

4.3.4 Formats for progress reporting

Progress reports should fulfill the requisites listed in the header 4.2.4 of this chapter. The main aspects to be fulfilled are:

Text:

The main text of the report should summarize the main aspects of the work carried out within the measuring Programme, especially:

a) Instrument set up dates and periods of data recording for the several stations. Record gaps occured.

b) Summary of the samples obtained as well as flow measurements performed by the operator, if any.

c) Description of the main events that happened and the water and sediment discharges observed and measured. Alternative estimates of sediment transport or error assessment, if possible. Assessment of the long-term representativity of the data obtained, based on the likely recurrence or frequency of the events recorded.

d) Discussion on the progress obtained relevant to the qualitative and quantitative knowledge of erosion and sediment transport processes.

e) Suggestions for improvement of the measuring Programme.

f) Management-oriented summary (see Conclusions).

Graphs:

• Hyetographs, hydrographs and sedigraphs of the more relevant measured events, comparing stations or seasons.

• Plots of the relationships between measured sediment discharges and water discharges that were used for the establishment of the sediment rating formulae.

Tables:

• Tables of daily total runoff and sediment transport with totals per month (these tables are for presenting information but the calculations are to be made in the 5 minute step spreadsheets).

• Summary of the main results obtained at the stations, per event, per years and for all the reported period of: precipitation, runoff, runoff coefficient, suspended sediment transport, dissolved sediment transport, mean erosion rate and chemical weathering rate averaged for all the catchment.

Enclosure:

Diskette with copies of the spreadsheet files used for the reported period.

Contents - Previous - Next