The pool of dissolved organic carbon (DOC) is a pivotal influencer in the ecology and biogeochemistry of inland water ecosystems, constituting a significant factor in the carbon budgets of terrestrial ecosystems. Employing a machine learning approach and leveraging a large curated dataset (AquaSat), this study delves into the importance of multiple spectral remote sensing data and spatiotemporal information in contributing to the variability of DOC concentrations in inland water bodies. The research underscores the critical role of spatial and temporal information in enhancing the accuracy of DOC predictions, revealing a remarkable 50% decrease in RMSE and MAE and a 110% improvement in R2in machine learning model performance when spatiotemporal covariates are incorporated. This marks a significant improvement over the reported accuracy on the same DOC dataset [1]. The dataset’s extensive coverage of inland water bodies provides valuable insights that are crucial for effective water quality monitoring, environmental management, and ecosystem health assessment.

The Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) provides essential 3-dimensional aerosol information in terms of height, which is crucial for understanding ongoing atmospheric processes related to aerosols. When CALIPSO observations are unavailable, a 3-dimensional cloud/aerosol categorization model based on past ground-level observations can play a significant role in predicting height-wise aerosol information. In this study, a machine learning-based algorithm for 3-dimensional aerosol-cloud prediction has been proposed. The algorithm utilizes CALIPSO lidar height-wise observations (Vertical Feature Flag), ground-level particulate matter (PM2.5) measurements with a diameter of ≤ 2.5μm, and meteorological variables (such as 2m-temperature, total precipitation, u- and v-components of wind, surface pressure) from the ERA-5 dataset. In addition, the model is utilised to illustrate its efficacy in 3-dimensional aerosol and cloud identification for Delhi, India’s capital city.

Excessive levels of carbon dioxide (CO2) in the atmosphere contributes to global temperature rise, and efforts are being made to limit this increase to ensure the safety of Earth’s inhabitants. Satellites like GOSAT-2 and OCO-2 provide global-scale monitoring of atmospheric CO2levels. However, cloud and aerosol occlusion result in missing data, and the spatial and temporal resolutions of these measurements are coarse. Addressing these limitations is crucial for leveraging satellite-based global CO2monitoring to identify CO2sources and sinks and understand their spatio-temporal evolution. In this study, we employ machine learning techniques to estimate column-averaged CO2(XCO2) from level 2 (L2) XCO2estimates obtained from the OCO-2 satellite, and daily XCO2data generated using Fixed Rank Krigging (FRK) at a spatial resolution of 10x10is used as the target variable. Meteorological variables, known to strongly influence XCO2distribution, are considered as covariates in the machine learning framework. To validate our estimates, we compare them with measurements from the Total Carbon Column Observing Network (TCCON) sensors. Additionally, we compare our estimates with those obtained from FRK and GEOS-L3 data. The validation against TCCON measurements and the comparison with existing data sources contribute to the evaluation and reliability of our approach.

We present a hybrid methodology using ensemble machine learning to estimate the surface soil moisture. We use satellite and physical model based soil moisture estimates along with satellite, soil and topographical data based derived features. We first process all these data and train multiple ML models using the above data with ground based soil moisture as target variable. Then, an ensemble model is built with the individual models. The approach is developed and tested on 107 in-situ sensor sites across Texas (USA) region over the period of 2015 - 2020. The gradient boosting models have the best overall performance. However, as expected, the ensemble approach improves upon the performance significantly compared to individual models with average correlation of 0.91 and RMSE of 0.025 (m3/m3). These results show the importance of ensemble approach and new features for improved soil moisture estimation.

In this article, we propose a method to solve the problem of optimal distribution of slots among sensor devices associated in wireless body area networks (WBANs). Modern healthcare is witnessing a paradigm shift due to the recent developments in Internet of Things (IoT), and WBAN is a fundamental enabling component of IoT-based healthcare. In a WBAN, sensors and actuators are implanted in patients (both on-body and in-body) connected to a hub to monitor health conditions ubiquitously and in real time. WBANs deal with heterogeneous sensors with diverse resource demands and constraints. Optimal allocation of data transmission slots among these heterogeneous sensors is a real challenge. Thus, in this article, we propose a cooperative game-theoretic approach, based on the Nash bargaining solution (NBS), for allocating slots for sensor devices in WBANs where the sensors communicate with each other following the IEEE 802.15.6 standard. We also compare the performance of our work with this standard and other relevant benchmarks to show the efficacy of the proposed solution. The proposed bargaining-based optimal slot sharing (BOSS) algorithm yields 26.68% better reliability and 38.07% better throughput, on an average, than the traditional IEEE 802.15.6 standard.

Wireless body area networks (WBANs) support various types of medical applications with heterogeneous requirements. Therefore, we need to use an efficient medium access control (MAC) protocol to ensure reliable data transmission. In this article, we propose a dynamic superframe structure-based MAC protocol extending the principles of the IEEE 802.15.6 standard. In this work, to allocate dedicated slots for each sensor device, a prioritized dedicated slot allocation mechanism using the criteria importance through intercriteria correlation (CRITIC) is proposed. With the help of this method, the priority value of sensor devices is calculated based on different sensors’ parameters. We compared the performance of our proposed work with standard IEEE 802.15.6 MAC and a few other MAC protocols. The simulation result shows that our proposed MAC protocol performed better in terms of energy efficiency and reliability, as well as reducing the packet drop probability. Results show that the reliability of data transmission increases over the IEEE 802.15.6 MAC protocol by more than 50%.

One of the major considerations in precision agriculture is optimizing fertilization to ensure maximum crop productivity. The principle behind this precision fertilization is to adjust the fertilizer inputs according to the properties of soils and crop stage such that crop yield is maximized and fertilizer loss is minimized. For this, one needs to understand the default soil nutrient status. This paper presents a feasible approach for developing high resolution spatial map (30m) of soil nutrients based on machine learning model with multiple sources covariates from climate, terrain, remote satellite etc. The covariate variables in different scales were chosen based on their association with soil nutrients and converted to construct field scale soil nutrients (pH, OM, EC, N, P & K) mapping models. The nutrients distribution is able to capture the spatial variability with moderate to high accuracy. This research is a methodological contribution to precision agriculture and lays the ground for precise application of fertilizers.

Sugarcane is a high biomass crop that requires large quantities of water for maximum yield. The study aims to estimate daily actual water consumption or crop evapotranspiration ( ETc) at field scales for practical applications with real data, remote sensing (RS) observation using novel methodologies (ML: machine learning and LSM: land surface model). Northeast Thailand is the study area chosen and three Sugarcane growing seasons (2016 to 2019) is the duration of the study. Similarities between the crop coefficient ( Kc) curve and a satellite-derived leaf area index (LAI) showed potential for estimation of Kc maps. A regression model has been developed to establish LAI vs Kc relation and used to derive daily Kc maps. With a view to computing daily reference evapotranspiration ( ETo) driven by weather, RS data, soil texture, land charac-teric etc, a high resolution LSM has been customized. The results shows ETc distribution low at initial and early development stages, while ETc tends to be high during grand growth and yield formation stages. Significant spatiotemporal variation has been observed across fields. Analysis of 19 fields for complete three seasons has been undergone with regard to yield response to water consumption and outcomes confirm standard yield reduction due to water stress. The daily ETc maps aided to demonstrate the variability of crop water use during growing season at field scales. Further, using ETc maps at the field scale in near real time, growers can supply optimal water to maximize yield, leading to water conservation in scale.

One of the major components of precision agriculture is the precision fertilization. The principle of precision fertilization is to adjust the fertilizer input according to the properties of soils at each location for the least waste and the highest production of yield. The paper presents a feasible approach for developing high resolution spatial map of soil nutrients based on k-nearest neighbor Convolutional Neural Network (CNN) model. Our CNN-based approach is appropriate for mapping soil nutrients because of its ability to predict soil nutrient value at less computational expense and real-time processing capabilities unlike traditional geostatistical methods. Using the field sampling database from Govt. of India, (Soil Health Card-SHC), soil nutrients (N, P, K & OC) map is generated at field scale (<; ha). The nutrients distribution is able to capture the spatial variability with high accuracy. This research is a methodological contribution to precision agriculture and lays the ground for precise application of fertilizers.

Soil nutrient estimation can be used as a key input to increase crop yield and agriculture fertilization. Due to the shortage of the ground measured spectrum technology and costliness to obtain hyperspectral images, multispectral remote sensing data is used to explore the soil nutrient content estimation. In this paper, we have used optical remote sensing data (Landsat-8 and Sentinel-2), terrain/climate data (precipitation, radiation, slope etc.) and ground truth value to estimate four nutrients: N, K, P, and OC for two districts of Maharashtra, India. We compared four linear and non-linear regression models: multiple linear regression (MLR), random forest regression (RFR), support vector machine for regression (SVR) and gradient boosting (GB) for estimation of NPK and OC. Comparative results suggest that, GB and RFR performed better than other models with sMAPE in range of 0.125-0.377 for all nutrients, which is better or comparable with literature reported accuracy [1]. Therefore, the approach has potential to generate high resolution (<; ha) soil nutrients map and can reduce soil sampling effort/cost.

Availability of soil moisture observations at a high spatial-temporal resolution is a prerequisite for various agricultural applications. This paper presents the comparison of SMAP (Analysis & Geophysical) and GLDAS soil moisture products versus a customized high resolution land surface model over a region representative of tropical regions located in Malaysia. The in-situ data over a nine month period is used to evaluate SMAP, GLDAS and customized LSM soil moisture. An overestimation of SMAP and GLDAS soil moisture products to in-situ data was noticed whereas customized LSM performed better with RMSE at surface 0.05 m3/m3(ubRMSE: 0.048 m3/m3) and rootzone 0.04 m3/m3(ubRMSE: 0.037 m3/m3) in this study. Both the SMAP products found similar and GLDAS is marginally better than SMAP. The result of this study is useful to support the continuous improvement of the SMAP soil moisture retrieval model, especially in tropical regions. The evaluation across the study region indicates that the customized LSM is able to capture spatial-temporal variation of soil moisture with better accuracy.

There have been several efforts to utilize satellite-based synthetic aperture radar (SAR) measurements to determine surface soil moisture conditions of agricultural regions. The results have been mixed since the relation between the SAR signal and surface soil moisture is confounded by variations in topographic features, surface roughness and vegetation density etc. We designed an experiment to investigate SAR based soil moisture retrieval using different machine learning techniques. In addition, a high resolution land surface model customized and deployed for generating soil moisture at 250m resolution using various static and dynamics input data.

Soil moisture and temperature are key inputs to several precision agricultural applications such as irrigation scheduling, identifying crop health, pest and disease prediction, yield and acreage estimation, etc. The existing remote sensing satellites based soil moisture products such as SMAP are of coarse resolution and physics based land surface model such as NL-DAS, GLDAS are also of coarse resolution as well as not available for real time applications. Keeping this in focus, we have customized high resolution land data assimilation system (HRLDAS) for India. The customization involve: (1) use of Global Data Assimilation System (GDAS) dataset for dynamic forcing fields, (2) ability to ingest local information about the soil characteristics (3) high resolution USGS land-cover and other static datasets, amongst others. In this paper, we present the performance of the customized model against SMAP soil moisture data and local sensors observations. The first results from the comparison shows a significantly reduced errors in model results. The RMSE for LSM generated outputs are less than 2%, whereas SMAP gives 4-5 % error for soil moisture.