Revision as of 10:46, 1 September 2006; view current revision
←Older revision | Newer revision→
Partly by R. Gommes
Understanding Water Balance Output
Examining the water balance output information generated.
When the water balance calculations have been finalized 5 output screens are presented. All 5 are explained below.
Summary information per station.
This window contains the summary output. In this example theoutput is for a maize crop at the station Kavimba in Botswana. The cycle length of the crop is 12 dekads. The total water requirement (both normal and actual) is 519 mm. The planting dekad is 34 (first dekad of December 1991. No irrigation is applied, but the bund height of 75 mm would allow 75 mm of water on top of the field. The crop coefficient of the bare soil at the start of the season is 0.19.
Water Balance details per station.
This is the detailed output of the water balance for a station. Parameters which are calculated: the total water requirements, surplus water, water deficit, the FAO water satisfaction index and, most important, the actual evapotranspiration ETA. These values are aggregated to crop phenological stage at the bottom. The "normal index" (last line) is the water satisfaction index which would be obtained under average conditions of rainfall and ETP, but all other input parameters remaining equal (planting date, % effective rain...). The other variables (columns) have the following meaning:
- NOR: the normal rainfall for the dekad (mm);
- ACT: the actual rainfall for that dekad and season (mm);
- WRK: the working rain.
- PET: (actual or normal, depending on which is available)
- KCR: the crop coefficient as calculated by the programme;
- WR: the water requirements in mm, the product of PET and KCR;
- AvW: the water available to the crop (WRK-WR), mm;
- SW: the soil moisture at the end of the dekad (mm);
- S/D: the water surplus or deficit (positive: surplus; negative: deficit).
- INDEX: the water satisfaction index.
Rangeland index per station.
The Rangeland-index is the classic FAO water satisfaction index computed for periods of 5 dekads, with average evapotranspiration kept at potential level (KCR=1) and an assumed WHC of 50 mm. Similar to moving averages, the value assigned to a dekad corresponds to the five-dekad period centred about that dekad. Thus 3 for dekads 1 to 5, 4 for dekads 2 to 6... The rangeland index, when plotted over time, assumes a smoothed aspect and is associated with crop phenology. Based on local experience, there is usually a rather good association between planting and the RI. For instance, in Tanzania, it was found that the planting of millet usually occurs when the rising RI curve crosses the 0.4 (40%) threshold. The RI is computed using actual rain. The normal crossing dates are also given. They are a convenient tool to decide if the current season is early or late. Also note that more than one crossing at the low levels (30 or 40%) usually indicates a false start of the rains.
Graph for one station.
A number of parameters shown in the previous tabs are represented graphically in this window. The parameters shown are:
- The actual rainfall for that dekad and season (mm);
- The water requirements in mm, the product of PET and KCR;
- The water satisfaction index.
- Irrigation. Is this case, no irrigation is applied.
Index distribution for all stations.
This graph gives an overview of the final index for all stations. In this case many stations have an index of 100% (no water stress).
Examining the detailed output files.
All output is written to 7 files, of which the first is in a plain text format. It can be read with any editor or word-processor and be easily printed.
(1) The text output file, ccWsyyOX.TXT (or cc2syyOX.TXT when the current year is post 1999) contains details of the water balance calculations, station by station and dekad by dekad. A sample TXT output is given as annex 7.
The other output files are in DAT format. They all contain the following variables, next to the station name:
Id, the crop identification number (see 3.1 above: maize=1,... ETP=8). ETP is a "dummy" crop with no phenology (the crop coefficient stays permanently at 1.0);
X, Y, Z are longitude, latitude, and elevation. As these parameters are given for reference only (they enter no calculations) they can be given as "missing" in the input files without any problem to FAOINDEX. But they will probably be needed in the further processing of the files and, therefore, it is good practice to use correct values.
The additional variables which occur in each of the output files are described hereunder together with the description of files (2) to (7).
(2) the summary output file ccWsyyXA.* (or cc2syyXA.*) lists the parameters which are likely to be used for further processing and analysis, in particular for crop forecasting. The remaining files are more "monitoring-oriented" as they contain sequences of several dekad values for display and inter-station comparisons (for instance with ADDATI). The variables listed in the summary are the following:
H and WHCi, the soil Water Holding Capacity, and its initial value, respectively;
E and EfR%, the ratio between effective and actual total and rain, in percent;
P, the planting dekad (1-36);
C, the length of the crop cycle (dekads);
TWR, total water requirements in mm;
Indx, IndxNor and IndxLast, FAO soil water satisfaction index, in % of requirements, respectively actual, average and "last" values. Actual corresponds to the value expected at the end of the cycle, average is the end-of-cycle index computed with average rain and PET, all other parameters being those from the crop file; the "last" value is that of the last dekad for which actual data are available ();
EXWT and DEFWT: excess and deficit water totalled over the growing cycle;
ETA: actual crop evapotranspiration (i.e. actual crop water consumption), a factor directly related with crop yield;
%Av, ("available") the percentage of actual rainfall data used for the calculations (the rest being assumed to be average)
Cr1 to Cr4 and Ncr1 to NCR4 (for "Crossing" and "Normal Crossing") indicate the dates (dekads) when the rangeland index (see (7) below) crosses the 0.4*PET line. The dekads given in the file can be used to map the beginning of the actual (current) season (Cr1 to Cr4) and the normal beginning (NCr1 to NCr4). In areas with only one rainfall peak, there is only one normal start, but there are two in bimodal rainfall areas. In marginal areas, where rainfall is very close to 40% of PET, there may be more than one start even in areas with unimodal rainfall. The actual starts may be more than the normal starts, particularly in the not so rare event of false starts when rain decreases again after a first start: in the field, farmers must frequently replant when this happens.
(3) ccWsyySA.* (or cc2syySA.*) contains the soil water data, computed for each dekad, in the rows of the file. Note that dekads are crop dekads (from 1 to cycle length). Values beyond cycle length are coded with the default value for missing data. D1,D2...D36 are the dekad numbers in crop dekads (i.e. planting dekad is 1 and the last values corresponds to C, the cycle length).
(4) ccWsyyTA.* has the same structure as file (3); it contains the estimated crop actual evapotranspiration for each dekad.
(5) ccWsyyIA.* list the values of the FAO water satisfaction index.
(6) Water surplus in mm (positive values) and deficits in mm (negative) constitute one of the methods to assess visually periods of drought or water logging (file ccWsyyUA.*)
(7) The "rangeland-index" (RI; file ccWRyyOA.* or cc2RyyOA.*) is the classic FAO water satisfaction index computed for periods of 5 dekads, with average evapotranspiration kept at potential level (KCR=1) and an assumed WHC of 50 mm. Similar to moving averages, the value assigned to a dekad corresponds to the five-dekad period centred about that dekad. Thus 3 for dekads 1 to 5, 4 for dekads 2 to 6... The rangeland index, when plotted over time, assumes a smoothed aspect (Figure 1) and is associated with crop phenology. Based on local experience, there is usually a rather good association between planting and the RI. For instance, in Tanzania, it was found that the planting of millet usually occurs when the rising RI curve crosses the 0.4 (40%) threshold.
The RI is computed using actual rain (not "working rain" as described under 2.3, i.e. future values are coded as missing and not taken as their normal).
The dekads when the RI curve passes the 30 to 70% thresholds is given in the text output file ccWsyyOX.TXT or cc2syyOX.TXT .
