Land degradation due to hydro-salinity in semi-arid regions Using gis and remote sensing

Unsupervised classification was performed, as the complete site information was not available. It involved various processes like image enhancement, creating composite files, using cluster/ iso-cluster modules etc. Various salinity indices were also applied based on the concept of spectral response of salt-affected soils. It is to note that spectral response in terms of digital no. (DN) of salt-affected soils is relatively higher than other categories in band-1 (B1) and band-3 (B3). The following two salinity indices namely S.I (Salinity Index) and NDSI (Normalized Differential Salinity Index) proposed by Tripathi et. al. (1997) were applied which relatively give good results in re-classification of salt-affected lands. NDSI is just the reverse of NDVI index for vegetation.

S.I = (Band 1 X Band 3) ½ (1)
NSDI = (Band 3 - Band 4) / (Band 3 + Band 4) (2)

The strongly visible salt-affected lands were then extracted from the final classified image to make a thematic map. The sum of all the extracted salt-affected soils was prepared, considering the rest of area as slightly saline to normal lands (Figure 4). Depth to watertable data for the year of 1990, 1993, and 1996 for more than 150 piezometer, reported by JIID (1997) along with their GPS locations, were used to see the spatial watertable positions in the area.

Using moving surface interpolation technique, initially the average watertable positions maps were prepared and analyzed for each year and then a map of "mean watertable positions" based on the three year results, was prepared. It is further grouped as waterlogged and non-waterlogged conditions, assuming an assumption of critical watertable depth of _ 200m for the area (Figure 5).

Reflectance variations of vegetation on the image are attributed to the different species of vegetation and their densities, which together provide evidence of shallow ground watertable conditions and saline agricultural areas. Favorable growth conditions prevail in regions where the watertable is situated below the area of influence of evapotranspiration, that is, within 10-m depth (Srivastava 1997). An indication of whether scanty vegetation in an area is due to high watertable depth or salinity can be investigated using Normalized Differential Vegetation Index (NDVI)

which easily grasp the state of vegetation. The NDVI for the area can be expressed as:

NDVI = (NIR Band _ Red Band) / (NIR Band + Red Band) (3)

The maps of irrigation canals and surface drainage network were digitized in the GIS environment. Some discrepancies of mismatching were found when overlaid on the geo-referenced image. These layers then were corrected with the help of topographic maps and image. It was presumed that the topographic sheets could be the most reliable source of information. Buffering analysis along canal and drain channels was performed for 1000m corridor to see the hydro-salinity impact on productivity in terms of vegetation vigor.

Result and Discussion
Waterlogging and Salinization The IRS composite image was systematically visually interpreted with the help of old salinity maps/ information developed in early 80's. It is revealed that salt-affected and waterlogged areas are clearly delineated compared with other land use features in FCC pellet. The most of the salt-affected lands were observed more pronounced near the drainage network as shown in white to light bluish tones. JIID (1997) and Casas (1995) reported similar results. This could be best