Estimation of water stress variability for a rice agriculture system from space-borne hyperion imagery

Leaf relative water content (LRWC) for paddy crop is a key component in evaluating crop growth. Present study aims to assess the variability of water stress in a rice agriculture system for summer/rabi and winter/kharif rice during pre-monsoon and monsoon period respectively. Amongst the examined indices, water band index (WBI), normalized water index-3 (NWI-3), normalized water index-4 (NWI-4) are considered to indicate water stress in paddy crop based on statistical performance of ground based hyperspectral data. Regression models are developed between each index and LRWC of paddy crop at three phases - specifically the vegetative, reproductive and ripening phases of paddy crop. The result showed that the reproductive phase of paddy crop is the critical growth phase for the aforementioned indices which followed different linear and nonlinear relationships. Moreover, mapping of LRWC of rice agriculture system from hyperspectral imagery is achieved by employing the best index models such as WBI, NWI-3 and NWI-4. The best estimates of LRWC at the reproductive phase for pre-monsoon (summer) rice agriculture system exhibited wide variation from 61.29 to 91.23% (WBI), 62.01 to 87.49 %(NWI-3) and 54.95 to 89.14% (NWI-4). On the other hand, the best VNIR index models (WBI, NWI-3 and NWI-4) that well predicted the LRWC of paddy crop varied from 60.21 to 90.45% (WBI), 58.58 to 85.67% (NWI-3) and 49.08 to 90.40% (NWI-4) for winter rice agriculture system. It is interesting to note that LRWC mapped from Hyperion image has a good agreement with the observed in-situ LRWC value (60.64-91.82%) obtained for the rice varieties in an Indian rice agriculture system. This study will provide essential information to assess the paddy field heterogeneity in a rice agriculture system.