|
|
|
Overview |
Crop Production |
Crop Pattern |
Crop Yield |
Irrigation |
Soil Management | Precision Farming |
Relevant Products |
Relevant Links
Performance evaluation of an irrigation project using satellite Remote Sensing GIS & GPS
How Spatial Information Technologies help in Performance Evaluation of an Irrigation Project?
We have in mind 4 contemporary technologies, namely, (i) satellite remote sensing (SRS), (ii) digital image processing (DIP), (iii) geographic information system (GIS), (iv) global positioning system (GPS) to apply to address this issue.
- Satellite Remote Sensing
Crop classification using multi-spectral data of IRS LISS is now well-established operational tool in India. Based on the crop calendar, optimal satellite datasets covering the entire crop season (say, for rabi crop season, one data set each month during November to March) are selected. Before classification of an image, few pre-processing steps, like geometric rectification of satellite data using ground control points (GCPs) and normalisataion of multi-sensor image data are followed.
- Digital Image Processing
Geometrically rectified multi-date satellite data is sequentially analysed with maximum likelihood classifier algorithm (in the present study, using Erdas Imagine ver. 8.3.1 in Silicon Graphics workstation platform) supported with ground-truth (training sets) collected during field visit. Various crops are identified and extracted sequentially at different stages of the classification process from the multi-date satellite data during the crop growth stage period. Classification accuracy with respect to training sets and total thematic map is ensured by close examination of error matrix (confusion matrix). Multi-date satellite data are also transformed to NDVI (Normalised Different Vegetation Index) values using the mathematical function of ratio of difference between reflected radiation in infrared and red band values by sum total of these two band values. NDVI represents the integral effect of various factors that influence crop production and is well accepted in India and elsewhere for crop yield forecast and agricultural drought monitoring.
- GIS Application
Usually, maps available with Command Area Development Authority (CADA) are of various formats and themes consisting of canal network, irrigation command boundary / administrative jurisdiction, place names, besides rivers / streams, reservoir / tanks, settlements, transportation networks, etc. A digital database is created of this information, which is co-registered with multi-date satellite data. When the crop classification maps and crop condition (NDVI) maps are available from satellite derived data, extraction of information within various spatial domains of the irrigation command can be done by overlaying digital database. GIS modules in Erdas software are used for these purposes.
- GPS Application
Satellited based GPS provides accurate geo-referenced (in terms of latitude & longitude) positional location on the ground. Stand-alone mode GPS instrument is used to determine the geo-reference position of CCE (Crop Cutting Experiment) plots in the crop fields, maintained by CADA for crop yield estimation. These CCE plots (usually 5 metre x 5 metre in size) are then identified onto the satellite image to extract NDVI values of the CCE plots. A regression model between NDVI and crop yield for the sample CCE plots is developed which is then extrapolated to the spatial domain(s) of the irrigation command for estimation of crop production.
Case Study: Nagarjunasagar Irrigation Command
Nagarjuna Sagar Left Canal (NSLC) command is designed for irrigation of 0.42 million hectare CCA in 3 districts in Andhra Pradesh State. The spatial domain of NSLC consists of (i) 3 irrigation zones which is further sub-divided into 32 irrigation blocks (iii) located in 3 districts and (iv) is supported by 114 canal distributaries. Two time domains and one crop season, namely, rabi crop season (November-March) of 1990-1991 and 1998-1999 are selected to identify progressive performance in NSLC command. The methodology to use SRS, DIP, GIS and GPS to transform primary data (satellite image / CADA maps & office data) to secondary information (performance matrix) is followed, as mentioned earlier. All spatial (maps) and attribute (crop types, crop area, cropping intensity, and crop yield) information are generated with reference to 2 time domains and 4 spatial domains. In the following statements, the portion ‘underline’ indicates the performance Lmatrix in the NSLC irrigation command as a whole. Performance matrix with respect to disaggregated spatial units are cited in Ref.3.
- Crop classification from IRS 1A LISS II and IRS – 1C/1D LISS III data consists of paddy, cotton, pulses, tobacco, surgarcane, chillies and orchard crops. These crops are grouped into ‘irrigated wet’ and ‘irrigated dry’, as operationally followed in this command. Inconsistency in cropping pattern during the two time domains is observed (Figure 1). Irrigation service being extensive in 1998-1999 compared to 1990-1991, irrigation potential utilized is higher by 55,000 hectare and cropping intensity has increased from 71% to 84% (Figure 2).
* Paper presented in : 5th Annual International Conference, “Map India 2002”, organised by CSDMS at New Delhi, 6-8 February 2002.
- A crop yield model for the major crop (paddy) has been developed by correlating satellite data based time composited NDVI values at CCE plot locations with the estimated yield in the CCE plots (Figure 3). The model provides crop yield in tonnes / hectare, which is an important performance indicator across space and time domains. The model is validated and is found to predict well at the larger disaggregated units. In the process of crop yield model development, satellite based GPS technique is used to determine the geo-referenced locational information of CCE plots in the satellite image. Although, irrigation service was extensive in 1998-99, paddy crop productivity (Figure 4) was found to be lower (4.00 tonne / hectare) compared to 1990-91 (4.55 tonne / hectare). However, these yield estimates are much above the national average from the irrigated croplands.
- Water release into the canals is utilized to calculate Water Utilisation Index (WUI). During 1990-91, area irrigated was much less for per unit of canal irrigation water supplied (65 hectare irrigated / million cu.m water supplied), compared to 1998-99, when better utilization of canal irrigation water was achieved (92 hectare irrigated / million cu.m water supplied) {Figure 5}.
- All the performance parameters,
therefore, indicate general improvement in system performance of
NSLC project.
Acknowledgement
Ministry of Water Resource (MoWR), Govt. of India commissioned the satellite remote sensing based evaluation study of 14 major irrigation projects. NRSA acknowledges with great appreciation about this progressive step taken by the MoWR. Project team (authors of this paper) in NRSA places on record their sincere thanks to the Administrator and engineers of the Nagarjunasagar canal command project for access to their data and to Dr.R.R. Navalgund, Director, NRSA and S.K. Bhan, Deputy Director (Applications), NRSA for encouragement to present this paper in “Map India 2002” Conference.
Reference
- Chakraborti, A.K. 1999. ‘Evaluation of Water Management in Irrigated Croplands” Proc. IT IS, Aurangabad.
- Rao, V.V., Chakraborti, A.K. 2000. “Water Balance Study and Conjuctive Water Use Planning in an Irrigation Canal Command : A Remote Sensing Perspective”. International Journal of Remote Sensing, Vol.21, No.17, P.3227-3238.
- Water Resources Group, NRSA. “Satellite
Remote Sensing based Evaluation Study of Nagarjunasagar Left Canal
Irrigation Command in Andhra Pradesh”. Project Report, Feb.2001,
168 P.
|
|
|