An attempt was made to quantify soil properties using hyperspectral remote-sensing techniques and machine-learning algorithms. In total, 100 soil samples representing various locations and soil-nutrient statuses were collected, and the samples were analyzed for soil pH, EC, soil organic carbon, available nitrogen (AN), available phosphorus (AP), and available potassium (AK) by following standard methods. Soil had a wide range of properties, i.e., pH varied from 5.62 to 8.49, EC varied from 0.08 to 1.78 dS/m, soil organic carbon varied from 0.23 to 0.94%, available nitrogen varied from 154 to 344 kg/ha, available phosphorus varied from 9.5 to 25.5 kg/ha, and available potassium varied from 131 to 747 kg/ha. The same set of soil samples were subjected to spectral reflectance measurement using SVC GER 1500 Spectroradiometer (spectral range: 350 to 1050 nm). The measured spectral signatures of various soils were organized for developing a spectral library and for deriving various spectral indices to correlate with soil properties to quantify the nutrients. The soil samples were partitioned into 60:40 ratios for training and validation, respectively. In order to select optimum bands (wavelength) from the soil spectra, we have employed metaheuristic algorithms i.e., Particle Swarm Optimization (PSO), Moth–Flame optimization (MFO), Flower Pollination Optimization (FPO), and Battle Royale Optimization (BRO) algorithm. Further partial least square regression (PLSR) was used to find the latent variable and to evaluate various algorithms for their performance in predicting soil properties. The results indicated that nutrients could be quantified from spectral reflectance measurement with fair to good accuracy through the Battle Royale Optimization technique with a $R^2$ value of 0.45, 0.32, 0.48, 0.21, 0.71, and 0.35 for pH, EC, soil organic carbon, available-N, available-P, and available-K, respectively.

Background: The Agricultural drought during kharif seasons of the year 2019 to 2023 in Tamil Nadu state in India was analyzed in using the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS). Monthly precipitation data from 1991 to 2023 were obtained for the study, with the main objective of evaluating the duration, spatial extent, severity and lag time of meteorological and agricultural drought in the study area. Methods: The Enhanced Vegetation Index (EVI) was generated by the Moderate Resolution Imaging Spectroradiometer (MODIS) data. The pixel reliability layer was introduced based on the calculation of the cloud coverage at the moment of image acquisition. The Standardized Precipitation Index SPI in one month time scale plays very significant role in any vulnerability studies for accurate prediction of any events. Result: In the present study, the EVI for Kharif season was considered for five years i.e., 2019-2023 and it was correlated with the SPI at various time scales. The correlation coefficient of SPI was 0.39 with 0.1% level of significance, in other words, EVI 2022 was well correlated with SPI-6. Also, the relation between EVI and precipitation with 9 months interval was also tested. The 12 months interval of precipitation have more stress on vegetation and it can negatively impact the agriculture activities leading to crop failures.

The study predicted the soil erosion map of the Coimbatore district, Western part of Tamil Nadu, India, to assess the potential of Digital Soil Mapping (DSM) against the empirical model, Revised Universal Soil Loss Equation (RUSLE). A total of 386 surface soil samples were collected and their properties used to predict soil erosion spatially. The Geographical Information System (GIS) enabled RUSLE factors to be derived spatially and integrated for annual soil loss estimation. The soil loss values calculated based on the RUSLE factors ranged from 0 to 88.58 t ha^–1 yr^–1. The spatial soil erosion was further predicted based on the SCORPAN model using the decision tree methodology to compare the suitability of the approaches in erosion mapping. The modelling tool, Cubist, was adopted to model the continuous soil erosion properties with 33 environmental covariates as independent variables. The LS factor, profile curvature, geology, mean annual rainfall, maximum air temperature, land use, minimum air temperature, elevation, slope and geomorphology covariates were the most influencing variables in predicting soil loss. The Cubist model’s rules are applied to the NLCD-Cubist classifier to generate a soil erosion map, and the annual soil loss ranged from 0 to 52.54 t ha^–1 yr^–1. The spatial pattern of soil erosion classes indicated that the Coimbatore district’s western part has a high erosion risk. Based on the validation sample points (n = 190), the DSM approach showed a higher concordance correlation (r = 0.857), whereas the RUSLE model showed less agreement (r = 0.632). From the concordance correlation values, it could be inferred that the decision tree algorithm performed better than the RUSLE erosion model. However, from the error map generated, it is advisable to have more soil observations with suitable environmental covariates to improve digital soil erosion mapping accuracy. The DSM approach is exploited well with point-based soil erosion loss information. The RUSLE model would perform well if the R, K, LS, C, and P factor maps were available at uniform scale and spatial resolution.

Large-scale mapping of soil resources can be crucial and indispensable for several of the managerial applications and policy implications. With machine learning models being the most utilized modeling technique for digital soil mapping (DSM), the implementation of model-based deep learning methods for spatial soil predictions is still under scrutiny. In this study, soil continuous (pH and OC) and categorical variables (order and suborder) were predicted using deep learning–multi layer perceptron (DL-MLP) and one-dimensional convolutional neural networks (1D-CNN) for the entire state of Tamil Nadu, India. For training the deep learning models, 27,098 profile observations (0–30 cm) were extracted from the generated soil database, considering soil series as the distinctive stratum. A total of 43 SCORPAN-based environmental covariates were considered, of which 37 covariates were retained after the recursive feature elimination (RFE) process. The validation and test results obtained for each of the soil attributes for both the algorithms were most comparable with the DL-MLP algorithm depicting the attributes’ most intricate spatial organization details, compared to the 1D-CNN model. Irrespective of the algorithms and datasets, the R² and RMSE values of the pH attribute ranged from 0.15 to 0.30 and 0.97 to 1.15, respectively. Similarly, the R² and RMSE of the OC attribute ranged from 0.20 to 0.39 and 0.38 to 0.42, respectively. Further, the overall accuracy (OA) of the order and suborder classification ranged from 39% to 67% and 35% to 64%, respectively. The explicit quantification of the covariate importance derived from the permutation feature importance implied that both the models tried to incorporate the covariate importance with respect to the genesis of the soil attribute under study. Such approaches of the deep learning models integrating soil–environmental relationships under limited parameterization and computing costs can serve as a baseline study, emphasizing opportunities in increasing the transferability and generalizability of the model while accounting for the associated environmental dependencies.

Magnaporthe oryzae, the source of the rice blast, is a serious threat to the world’s rice supply, particularly in areas like Tamil Nadu, India. In this study, weather-based models were developed based on count time series and machine learning techniques like INGARCHX, Artificial Neural Networks (ANNs), and Support Vector Regression (SVR), to forecast the incidence of rice blast disease. Between 2015 and 2023, information on rice blast occurrence was gathered weekly from three locations (Thanjavur, Tirunelveli, and Coimbatore), together with relevant meteorological data like temperature, humidity, rainfall, sunshine, evaporation, and sun radiation. The associations between the occurrence of rice blast and environmental factors were investigated using stepwise regression analysis, descriptive statistics, and correlation. Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) were used to assess the model’s prediction ability. The best prediction accuracy was given by the ANN, which outperformed SVR and INGARCHX in every location, according to the results. The complicated and non-linear relationships between meteorological variables and disease incidence were well-represented by the ANN model. The Diebold–Mariano test further demonstrated that ANNs are more predictive than other models. This work shows how machine learning algorithms can improve the prediction of rice blast, offering vital information for early disease management. The application of these models can help farmers make timely decisions to minimize crop losses. The findings suggest that machine learning models offer promising potential for accurate disease forecasting and improved rice management.