Remote Sensing and GIS Technologies Developed

Land use land cover mapping and mapping of soil resources in various watersheds in Tamil Nadu were done with remote sensing data. Soil mapping at 1:50,000 scale using PAN merged LISS-III data was completed for entire Tamil Nadu. Digital format is available in this department and shred with user organisations.

Mapping of land degradation, gypsum mined area and cement factory pollution in TamilNadu was done using multi-season IRS LISS-III satellite data at 1:50,000 scale. Among the degradation types the sheet erosion by water contribute to 67.6 per cent of the degradation followed by Alkalinity (slight) 7 percent, acidity (moderate) 4 percent and barren rock 4 percent in the state of Tamil Nadu. At district level the sheet erosion by water is the major land degradation factor observed in almost all the district. This was followed by acidity (moderate) in districts of Cuddalore, Nilgiris and Ariyalur and sodic (slight to moderate) in Cuddalore, Madurai, Perambalur, Tirunelveli and Toothukudi districts. Presence of barren rock / stony wastes are observed in Kanyakumari, Salem, Namakkal, Kancheepuram and Theni districts.

Village level base map containing detailed natural and manmade features and survey boundaries were prepared for 512 villages covering 10 blocks. Village level soil maps containing detailed soil resources information at farm level were prepared. Analytical data for soil characterization and soil database were generated. Village level GIS database on soil resource for the 512 villages in Thirumanur, Annur, Pappireddipatti, Ottanchatram, Uthangarai, Rasipuram, Perambalur, R.S.Mangalam, Veerapandi and Gingee blocks of Ariyalur, Coimbatore, Dharmapuri, Dindugal, Krishnagiri, Namakkal, Perambalur, Ramnad, Salem and Villupuram districts respectively were generated.

Hyperspectral remote sensing technique for identification of crop nutrient deficiency especially nitrogen in Maize was attempted with the collaboration of National Remote Sensing Centre, ISRO, Hyderabad. This technique was tried under field condition by creating varying levels of nitrogen deficiency. The results suggest that this needs to be studied thoroughly with repeated laboratory and field experiments.

Village level soil maps containing detailed information on soil resources at farm level were prepared for Sarkar Samakulam, Valapadi, West Arani and Sathankulam Blocks in Coimbatore, Salem, Tiruvannamalai and Thoothukudi districts respectively. Analytical data for soil characterization and soil database were generated. Village level GIS database on soil resource for the 102 villages in Sarkar Samakulam, Valapadi, West Arani and Sathankulam Blocks were generated.

Soil moisture content was estimated using SAR data from RISAT. The results of 1.12 to 28.43 per cent obtained from satellite data recorded higher correlation  with the gravimetric estimation with values from 1.20 per cent to 36.20 per cent in both linear function and polynomial function with a r² value of 0.631 and 0.735, respectively. This study of soil moisture using microwave RS data (RISAT 1) indicated a good correlation with the backscatter co-efficient for the study area.

The spectral reflectance curve of cotton plants damaged by pests namely thrips, leafhopper and aphids were different from that of the healthy plants. In general, in the damaged plants, there was a decrease in near infra red (NIR) reflectance (770 to 860 nm) while the blue (400 to 500 nm), green (520 to 590 nm) and red reflectances (620 to 680 nm) increased compared to undamaged plants. The per cent sensitivity of red band to thrips damage was the highest when compared to blue, green and NIR on most of the days of observation in both the variety and hybrid studied. In the sensitivity curve there was an increase in blue followed by a decrease in green, then an increase in red and finally decrease in NIR region on all days of observation. The per cent sensitivity of red reflectance to sucking pest damage was the highest when compared to other bands on most of the days of observation in both the variety and hybrid studied.

Based on the field studies it was found that the spectral reflectance curve of rice hills damaged by pests namely Leaf folder (LF) and Yellow stem borer (YSB) was different from that of the healthy hills. In general, in the damaged hills, there was a decrease in near infra red (NIR) reflectance (770 to 860 nm) while the green (520 to 590 nm) and red reflectances (620 to 680 nm) increased compared to undamaged hills. There was a significant negative correlation between LF damage and normalized difference vegetation index (NDVI), simple ratio (SR) and green red vegetation index (GRVI) values on all days of observation (R² values above 0.70). Similarly negativee correlation was also observed between YSB damage and NDVI, SR and GRVI values on most of the days of observation (R² values above 0.70 on most of the days).

Based on the results obtained on sensitive bands for the damages caused by the two pests studied in field on majority of the days of observation, the following Damage Specific Vegetation Index (DSVI) is suggested for estimation of damage levels of LF and YSB which was found to have better correlation and sensitivity than other indices. The regression equations to predict the damage by all the three pests have been worked out based on DSVI.

Coimbatore has largest area under coconut among all districts of Tamil Nadu, followed by Tiruppur, Thanjavur and Dindigul.  In terms of percentage of coconut area to the total geographical area of the district, Tiruppur, leads the list, followed by Kanyakumari, Coimbatore and Thanjavur.

The soil data was linked with coconut map to generate soil related limitations which will affect the growth and productivity of coconut. The soil limitations and their area extent for various districts were generated. Among the various limitations the deficiency of micronutrients particularly Iron and Zinc accounts for 42.5 % of the total coconut areas followed by imbalance between calcium, magnesium and potassium (42.5 %) and calcareousness (42.1 %).

The results indicated that the Hyper-spectral data and derived vegetation index could be used to discriminate the crops, varieties, growth stages of crops. Pure crop of rice could be discriminated from weed mix up from spectral data. There is a clear discrimination of nutrient levels applied to rice crops and yield levels could also be discriminated. In Cotton crop varieties and Bt hybrid, species of cotton were also discriminated from both spectral data and vegetation index.

Cosmoskymed, TerraSAR-X and RISAT imageries with very good accuracy .Start of the Season and Peak of the season Maps and Rice yield map were generated.

SAR based Remote sensing products in integration with ORYZA model, were used to deriving crop variables such as LAI, Relative growth rate for estimating yields on a spatial scale through an interface.Yield Simulation accuracy of more than 87% at district, block and farm level from the study means that simulated yield matched observed yield perfectly indicating the suitability of these products for policy decisions

WRSI could effectively be computed from freely available remote sensing data for ascertaining crop water stress at regional level. The risk zones identified in each district for growing rainfed crops like Cotton, Sorghum, Maize and Groundnut can become a guide to plan for avoiding water stress and for better irrigation management.

The lacuna in the spatial availability of in-situ meteorological data for water management could be overcome by the use of open source remote sensing data.

Generated spectral reflectance curve of maize. Wavebands for nitrogen stress detection at 60 DAS: 545 nm (green), 660 nm (red) and 800 nm (NIR). 20 most effective wave bands which could discriminate treatments at all the stages were identified using STEPDISC procedure of SAS 9.0 viz., 716, 777, 781, 844, 1031, 1061 and 1073 nm at Infrared and SWIR regions along with two wavebands of 543.2 and 557.9 nm at green region.

The weighted combination of remote sensing drought indices of scaled LST, scaled TRMM, and scaled NDVI; scaled LST, scaled TRMM and scaled NDWI6; scaled LST, scaled TRMM and NDWI7 were identified as the optimum remote sensing-based drought indices that can be used for agricultural drought monitoring.

The methodology refined by this study can provide opportunity to further investigate satellite remote sensing, including advanced multi-sensor data assimilation to drought monitoring and prediction.

Using Spectro-radiometer GER 1500, spectral reflectance values of degraded lands was collected at different locations of AEZ 4.1 of Tamil Nadu uplands. Among the locations, the degraded status of lands was found to be of three different categories, viz., slight, moderate and severe. It was observed that the ratio of reflectance at 760 and 675 nm (R_760/675 ratio) has given a better discrimination between various types of degraded lands. R_760/675 values for slight and moderate degraded values were found to be 1.28 and 2.05 respectively. Hence, this ratio can be used for differentiating slight and moderately degraded lands

Flood map was generated using Sentinel 1A SAR data and area statisticts of Cuddalore District was generated. Groundtruth collection was also made to verify the flooded area through smartphone based applications. Parangipettai, Kurinchipadi, MelBhuvanagiri, Kumaratchi and Keerapalayam are the major blocks with inundated standing water as on 12th November, 2015.The products were generated within two days and shred with the State Government for relief measures.

The project was aimed at generating farm level nutrient balance data of 5 Agro-climatic zones viz., North Eastern, North Western, Western, Cauvery Delta and Southern Zones of Tamil Nadu. The results indicated that Negative nutrient balance especially in nitrogen and potassium were recorded In all the zones irrespective of farm size and management types. The findings were shared with Scientists and Extension workers working in these zones for formulating research agenda and effective technology transfer.

Based on Munsell Soil Colour measurements the samples falls in 5 Hue categories viz., 2.5YR, 5YR, 7.5YR and 10YR with varying value and chroma. The spectral signature of these soil samples indicates observable difference among different coloured soils in the visible as well as in IR region. The comparision of Munsell Soil colour value with Spectral signature are presented in Figure 1a,1b&1c. The darker the colour, lower is the quantity of energy or light reflected and vice versa. Thus, the Spectral measurement technique is able to differentiate soil colours as that of Munsell soil colour chart. The high reflection is also due to light coloured soil as well as fine soil texture and coarse soil texture would absorb more light. The advantage of Spectral measurements is observation in near infrared wavelength in addition to visible region. Thus, enables easy differentiation of different coloured soils.