Integrating effective drought index (EDI) and remote sensing derived parameters for agricultural drought assessment and prediction in Bundelkhand region of India
Padhee, S. K.
Nikam, B. R.
Aggarwal, S. P.
Drought is an extreme condition due to moisture deficiency and has adverse effect on society. Agricultural drought occurs when
restraining soil moisture produces serious crop stress and affects the crop productivity. The soil moisture regime of rain-fed agriculture
and irrigated agriculture behaves differently on both temporal and spatial scale, which means the impact of meteorologically and/or
hydrological induced agriculture drought will be different in rain-fed and irrigated areas. However, there is a lack of agricultural
drought assessment system in Indian conditions, which considers irrigated and rain-fed agriculture spheres as separate entities. On the
other hand recent advancements in the field of earth observation through different satellite based remote sensing have provided
researchers a continuous monitoring of soil moisture, land surface temperature and vegetation indices at global scale, which can aid in
agricultural drought assessment/monitoring. Keeping this in mind, the present study has been envisaged with the objective to develop
agricultural drought assessment and prediction technique by spatially and temporally assimilating effective drought index (EDI) with
remote sensing derived parameters. The proposed technique takes in to account the difference in response of rain-fed and irrigated
agricultural system towards agricultural drought in the Bundelkhand region (The study area).
The key idea was to achieve the goal by utilizing the integrated scenarios from meteorological observations and soil moisture
distribution. EDI condition maps were prepared from daily precipitation data recorded by Indian Meteorological Department (IMD),
distributed within the study area. With the aid of frequent MODIS products viz. vegetation indices (VIs), and land surface temperature
(LST), the coarse resolution soil moisture product from European Space Agency (ESA) were downscaled using linking model based on
Triangle method to a finer resolution soil moisture product. EDI and spatially downscaled soil moisture products were later used with
MODIS 16 days NDVI product as key elements to assess and predict agricultural drought in irrigated and rain-fed agricultural systems
in Bundelkhand region of India. Meteorological drought, soil moisture deficiency and NDVI degradation were inhabited for each and
every pixel of the image in GIS environment, for agricultural impact assessment at a 16 day temporal scale for Rabi seasons (October–April) between years 2000 to 2009. Based on the statistical analysis, good correlations were found among the parameters EDI and soil
moisture anomaly; NDVI anomaly and soil moisture anomaly lagged to 16 days and these results were exploited for the development
of a linear prediction model. The predictive capability of the developed model was validated on the basis of spatial distribution of
predicted NDVI which was compared with MODIS NDVI product in the beginning of preceding Rabi season (Oct–Dec of 2010).The
predictions of the model were based on future meteorological data (year 2010) and were found to be yielding good results. The
developed model have good predictive capability based on future meteorological data (rainfall data) availability, which enhances its
utility in analyzing future Agricultural conditions if meteorological data is available.