Remote sensing approaches based on VIS-NIR spectroscopy can be used for getting near real-time information about soil fertility. However, the main challenge limiting the application of spectroscopy in soil fertility evaluation is finding suitable data pre-processing and calibration strategies. We have compared various pre-processing techniques using the reflectance spectra obtained from AVIRIS-NG hyperspectral images, for quantification of organic carbon (OC), available phosphorus (P) and available potassium (K) in the surface soils of Surendranagar area (Western parts of India) and Raichur (Southern parts of India). Surface (0 - 0.15 m) soil samples were collected from these two areas synchronously with the dates of the AVIRIS-NG campaign. The soil samples were air dried, sieved to <2 mm, and analyzed for OC, P, and K using standard methods. The AVIRIS spectra (spectral range of 380-2500 nm with an interval of 5 nm) corresponding to soil sampling points were extracted. The pre-processing steps were used in the order: Continuum Removal (Yes/No), Moving Window Abstraction (Yes/No), No transformation or Euclidean Normalization or Standard Normal Variate (SNV), No transformation or Savitsky-Golay (SG) first-order smoothing, and No transformation or first derivative OR second derivative. We have used the partial least squares regression (PLSR) to calibrate the model from pre-processed spectra. The PLSR with Continuum Removal, SNV, SG first-order smoothing, and first derivative was selected as the best algorithm for estimating soil properties from the Western parts of India, and the corresponding R2 were 0.77 for OC, 0.79 for P and 0.83 for K (RMSE <0.3 for all the parameters). The PLSR with Moving Window Abstraction, SG first-order smoothing, and second derivative were selected as the best algorithm for estimating soil properties from the Southern parts of India, and the corresponding R2 were 0.54 for OC, 0.49 for P and 0.56 for K (RMSE <0.3 for all the parameters). These results suggest that the optimization of AVIRIS spectra using various pre-processing techniques and modeling approaches is required for rapid and non-destructive assessment and monitoring of soil health for precision agriculture.

Regional crop production estimates are important in both public and private sectors to ensure the adequacy of a food supply and aid policymakers and farmers in managing harvest, storage, import/export, transportation, and anticipate market fluctuations. Food security will be progressively challenged by population growth and climate change. Thus, the prediction of accurate regional crop yield is essential for national food security and the sustainable development of the Indian agriculture sector. In this study, we have selected Punjab, the highest wheat yielding state in India. The district-wise wheat yield data were available for the year 2000 – 2019. We have used several covariates for crop health viz. normalized difference vegetation index (NDVI), leaf area index (LAI), fraction of absorbed photosynthetically active radiation (fAPAR); meteorological indicators viz. land surface temperature (LST), and evapotranspiration (ET); and surface characteristics viz. protrusion coefficient (PC). These indicators were generated at 250 m spatial resolution from the MODIS data using Google Earth Engine. The whole data was divided into two groups for training (2000 – 2009, 2011, 2013, 2014, 2016 - 2019) and testing (2010, 2012, 2015), which were randomly selected. This study uses the random forest (RF) regression method to create a wheat yield prediction model. We created several combinations of covariates and found that fAPAR and ET are highly correlated with NDVI and do not have much influence on the model’s prediction accuracy. Hence, only four out of six covariates were selected for final training. The coefficient of determination between district-level yield vs. (NDVI/LAI/PC/LST) was 0.37/0.31/0.15/0.13 respectively. We used randomized search cross-validation as well as grid search cross-validation for hyper-parameter tuning. Furthermore, we used mean absolute error (MAE) and accuracy as quality metrics. The MAE for training was 0.1870 t/Ha with 95.81% accuracy, whereas the MAE on test data was obtained as 0.4293 t/Ha with 90.02% accuracy. The results of this study are within acceptable error limits of the published research articles. Overall, this study demonstrates that covariates derived from coarse resolution satellite data can predict district-level crop yield with reasonable accuracy.

We assessed the effect of temperature and tropospheric concentration of Nitrogen dioxide (NO2) and Carbon monoxide (CO) before and during the COVID 19 curfew/lockdown period (22 March-19 May 2020) on mortality due to COVID-19 in Indian Punjab. Time series daily data of TROPOspheric Monitoring Instrument on board Sentinel-5 Precursor was used to study the spatio-temporal changes in NO2 and CO from 15 March to 19 May 2020. Visible Infrared Imaging Radiometer Suite onboard Suomi NPP satellite was used to detect the active fires (from 15 April to 19 May 2020) due to crop residue burning. The COVID-19 mortality was calculated from number of deaths relative to number of cases preceding two weeks of deaths due to COVID-19 virus. The weekly (15-21 March 2020) averaged tropospheric concentration of NO2 and CO was higher before the first day of curfew (22 March 2020). The concentration of NO2 decreased during the curfew period followed by increasing concentration from 11 April to 19 May 2020 and the concentration of CO increased from 19 April to 19 May 2020. There was a continuous increase in daily air temperature from 15 March to 19 May 2020. Mortality due to COVID-19 virus was significantly negatively correlated with NO2, CO and temperature. These results show that increasing temperature and concentration of NO2 and CO decrease COVID-19 mortality, but further studies should be conducted in different continents of the world to verify the impact of NO2, CO and temperature on COVID-19 mortality.