Search

Accurate oil palm yield prediction is necessary to sustain oil palm production for food security and economic return. However, there are limited studies on comprehensive mapping and accurate oil palm yield prediction using advanced machine learning algorithms. Using multi-temporal remote sensing data, this paper proposed a new approach to predict oil palm yield based on the normalized difference vegetation index (NDVI) and ensemble machine learning algorithm. ReliefF algorithm with linear projection was employed to select the best combination of spectral indices in oil palm discrimination. Oil palm land cover was classified using random forest (RF) and modified AdaBoost algorithms. A time-series approach known as walk-forward validation was firstly introduced to train the model using the 2016-2019 data and the one-step prediction was performed for 2020 using RF and AdaBoost. Result of the study revealed that the RF model (RMSE=0.384; MSE=0.148; MAE=0.147) outperformed the AdaBoost model (RMSE=0.410; MSE=0.168; MAE=0.176). Our research has demonstrated the value of detailed mapping and subsequent yield prediction by developing a novel approach utilising time-series satellite imagery, ensemble machine learning, and NDVI, which will assist decision-makers in managing their practices related to oil palm. [ABSTRACT FROM AUTHOR]

This study compares the relationship between different NDVI (Normalized Difference Vegetation Index), the NDVI of AVHRR (Advanced Very High Resolution Radiometer) (NDVIa), the NDVI of MODIS (Moderate Resolution Imaging Spectrometer) (NDVIm), and the NDVI of VIRR (Visible and Infrared Radiometer) (NDVIv), and found that there is a significant correlation between the NDVIa and the NDVIm, and between the NDVIv and the NDVIa, the relationship between the three is NDVIv < NDVIa < NDVIm. Machine learning is an important method in artificial intelligence. It can solve some complex problems through algorithms. This research uses linear regression algorithm in machine learning to construct the Fengyun Satellite NDVI correction method. By constructing a linear regression model, the NDVI value of Fengyun Satellite VIRR is corrected to a level that is basically the same as NDVIm. The corrected correlation coefficients (R2) were significantly improved, and the corrected correlation coefficients were significantly improved, and the confidence levels were all significant correlations less than 0.01. It is proved that the corrected normalized vegetation index of Fengyun Satellite has significantly improved accuracy and product quality compared with the normalized vegetation index of MODIS. [ABSTRACT FROM AUTHOR]

As one of the earliest remote sensing indices, the Normalized Difference Vegetation Index (NDVI) has been employed extensively for vegetation research. However, despite an abundance of NDVI review articles, these studies are predominantly limited to either one subject area or one area, with systematic NDVI reviews being relatively rare. Bibliometrics is a useful method of analyzing scientific literature that has been widely used in many disciplines; however, it has not yet been applied to comprehensively analyze NDVI research. Therefore, we used bibliometrics and scientific mapping methods to analyze citation data retrieved from the Web of Science during 1985–2021 with NDVI as the topic. According to the analysis results, the amount of NDVI research increased exponentially during the study period, and the related research fields became increasingly varied. Moreover, a greater number of satellite and aerial remote sensing platforms resulted in more diverse NDVI data sources. In future, machine learning methods and cloud computing platforms led by Google Earth Engine will substantially improve the accuracy and production efficiency of NDVI data products for more effective global research. [ABSTRACT FROM AUTHOR]

The normalized difference vegetation index (NDVI) is a key parameter in precision agriculture. It has been used globally since the 1970s as a proxy to monitor crop growth and correlates to the crop coefficient (Kc), leaf area index (LAI), crop cover, and more. Yet, it is susceptible to clouds and other atmospheric conditions that might alter the crop's real NDVI value. Synthetic Aperture Radar (SAR), on the other hand, can penetrate clouds and is hardly affected by atmospheric conditions, but it is sensitive to the physical structure of the crop and therefore does not give a direct indication of the NDVI. Several SAR indices and methods have been suggested to estimate NDVIs via SAR; however, they tend to work for local spatial and temporal conditions and do not work well globally. This is because they are not flexible enough to capture the changing NDVI–SAR relationship throughout the crop-growing season. This study suggests a new method for converting Sentinel-1 to NDVIs for Agricultural Fields (SNAF) by utilizing a hyperlocal machine learning approach. This method generates multiple on-the-fly disposal field- and time-specific models for every available Sentinel-1 image across 2021. Each model learns the field-specific NDVI (from Sentinel-2 and Landsat-8) –SAR (Sentinel-1) relationship based on recent NDVI and SAR time series and consequently estimates the optimal NDVI value from the current SAR image. The SNAF was tested on 548 commercial fields from 18 countries with 28 crop types and, based on 6880 paired NDVI–SAR images, achieved an RMSE, bias, and R2 of 0.06, 0.00, and 0.92, respectively. The outcome of this study aspires to a persistent seamless stream of NDVI values, regardless of the atmospheric conditions, illumination, or local conditions, which can assist in agricultural decision making. [ABSTRACT FROM AUTHOR]

Land-use and land-cover (LULC) mapping in the complex area is a challenging task due to the mixed vegetation patterns, and rough mountains with fast-flowing rivers. In Vietnam, LULC update is not frequently. In this study, we applied a supervised machine learning (Random forest—RF) approach to mapping LULC in Thanh Hoa province, Vietnam from 2011 to 2015 utilizing multitemporal Normalized Difference Vegetation Index (NDVI) data from MODIS, combined with topographic features. Random forest classification (RFC) reached a total prediction accuracy of 91% and Kappa coefficient (K) of 0.89 across eight LULCs. Besides, the results showed that the features extracted from time-series NDVI comprising the mean of yearly NDVI, the sum of NDVI, and the topography were the important variables controlling the LULC classification. For similar studies on the distribution of LULC, the method proposed in this study could be helpful. [ABSTRACT FROM AUTHOR]

[Display omitted] • Validated and compared the effects of multiple spatial and temporal fusion methods on annual-scale NDVI reconstruction. • Proposed a new idea to improve the accuracy of traditional spatial and temporal fusion methods. • Analyzed the spatial and temporal changes of vegetation in Danjiang River Basin from 2001 to 2021. • Analyzed the influence of different driving factors on vegetation growth in Danjiang River Basin. The Normalized Difference Vegetation Index (NDVI) is an essential metric in vegetation monitoring for remote sensing applications. While there are numerous long-term low-resolution NDVI datasets available there remains an unmet need for high-resolution NDVI reconstructions over extended time frames. Existing research has not comprehensively assessed the efficacy of various temporal fusion techniques for NDVI reconstruction at large regional scales. Traditional Spatiotemporal Image Fusion (TSTIF) methods often suffer from limited fusion accuracy due to input data quality constraints. To address these limitations this study introduces an innovative Deep Learning-Enhanced Spatiotemporal Fusion Method. Deep learning algorithms are employed to refine the spatial resolution of input data thereby facilitating the separation of complex image elements located at feature boundaries. This improved approach significantly enhances fusion accuracy as validated against five established TSTIF techniques through empirical analysis. As a case study we generate long-term Fractional Vegetation Cover (FVC) datasets to investigate the ecological dynamics of the Danjiang River Basin. Our findings reveal substantial gains in NDVI reconstruction accuracy through the incorporation of deep learning into traditional fusion techniques. Among the methods tested the STRUM algorithm showed the greatest improvement with its R2 value increasing from 0.872 to 0.894 and a consistent reduction in Root Mean Square Error (RMSE). The FSDAF technique emerged as the most effective boasting an R2 value of 0.953 and an RMSE of 0.012. These enhanced NDVI datasets provide more accurate representations of the basin's vegetation cover thus enriching long-term observational archives and facilitating further quantitative remote sensing analyses. Moreover we report a 21-year mean FVC of 0.702 for the Danjiang River Basin characterized by a "north-high south-low" distribution that varies seasonally. Improvements in vegetation cover were observed in Xichuan and Neixiang counties while environmental degradation is evident in the broader Danjiang River watershed. Key factors influencing these spatial variations include elevation and soil composition in addition to climate variables such as rainfall and temperature. Infrastructure projects like the South-to-North Water Diversion Project and targeted ecological conservation policies have also positively impacted vegetation within the basin. [ABSTRACT FROM AUTHOR]

• NDVI has been estimated from sRGB co-registered images. • NDVI was estimated using a shallow-regressive neural network performing a pixel-to-pixel regression. • Calibrated RGB images have been used obtaining a correlation with NDVI in validation r = 0.91. • The model was applied to a dataset acquired with a different non-co-registered obtaining a r = 0.71. • The obtained model is ready to be applicable being available as supplementary material (GitHub). The Normalized Difference Vegetation Index (NDVI) is the most common index to measure vegetation in agriculture and create classified prescription maps used for several purposes. It is obtained as the ratio between information from the visible and near infrared spectral bands. The calculation of NDVI often uses very expensive hyperspectral and multispectral optical sensors. This study aims to develop an efficient model to extract NDVI data from RGB images. During the work were acquired images, of different plant species at different vegetation status, through a snapshot hyperspectral camera (Specim IQ) featuring natively superimposed RGB sensor whose images were calibrated a posteriori (sRGB). NDVI was predicted from sRGB images using a Shallow-regressive neural network that performs a pixel-pixel regression. The model has been then tested on around 1000 drone images acquired by a 6x Sentera sensor to test the efficiency of the applied model. The model has been used to estimate the numerical value of NDVI and the classes produced by a clustering method (k -means). The results of this work showed that the calibrated images had a very high correlation with the NDVI in validation (r = 0.91), maintaining good performances (r = 0.71) when applied to a set of data acquired with a different non-co-registered sensor. Therefore, for the first time this work has shown that the health of vegetation through the NDVI could be calculated using unexpensive RGB device adopting a pixel-pixel regression AI approach. The approach shows its importance especially for small-sized farms where profit would not allow for the budget to access multispectral cameras and heavy carriers (i.e., UAV). [ABSTRACT FROM AUTHOR]

The growth of precision agriculture has allowed farmers access to more data and greater efficiency for their farms. With consistently tight profit margins, farmers need ways to take advantage of the advancement of technology to lower their costs or increase their revenue. One area where these advancements can prove beneficial are in the measurement of vegetation indices such as the Normalized Difference Vegetation Index (NDVI) and Normalized Difference Red Edge Index (NDRE). Color maps representing these vegetation indices can be used to identify problem areas, plant health, or even places where spot applications are needed. These color maps help farmers to visualize these areas. Currently, a multi-thousand dollar multispectral camera, typically attached to an Unmanned Aerial Vehicle (UAV) during flight, is required for measuring these indices. This makes obtaining NDVI and NDRE somewhat cost prohibitive for most farmers. This work demonstrates a solution to this cost issue. The solution involves the use of a conditional Generative Adversarial Network known as Pix2Pix. By using Pix2Pix along with training data from UAV flights of corn, soybeans, and cotton, this paper highlights the potential for predicting comparable NDVI and NDRE with a low-cost Red-Green-Blue (RGB) camera. This paper proposes and assesses a cost-efficient method that can comparably predict these vegetation indices, resulting in cost-savings in the range of $5000 per UAV system. • Pix2Pix used with aerial crop imagery to estimate vegetation indices. • Deep Learning for estimation of vegetation indices from standard imagery. • Using aerial crop imagery to train a network capable of producing accurate vegetation index estimates. • Leveraging Artificial Intelligence for cost savings. • NDVI estimation with Pix2Pix. [ABSTRACT FROM AUTHOR]

As a link for energy transfer between the land and atmosphere in the terrestrial ecosystem, karst vegetation plays an important role. Karst vegetation is not only affected by environmental factors but also by intense human activities. The nonlinear characteristics of vegetation growth are induced by the interaction mechanism of these factors. Previous studies of this relationship were not comprehensive, and it is necessary to further explore it using a suitable method. In this study, we selected climate, human activities, topography, and soil texture as the response factors; a nonlinear relationship model between the karst normalized difference vegetation index (NDVI) and these factors was established by applying a back propagation neural network (BPNN), a radial basis function neural network (RBFNN), the random forest (RF) algorithm, and support vector regression (SVR); and then, the karst NDVI was predicted. The coefficient of determination (R2), mean square error (MSE), root mean square error (RMSE), and mean absolute percentage error (MAPE) of the obtained results were calculated, and the mean R2 values of the BPNN, RBFNN, RF, and SVR models were determined to be 0.77, 0.86, 0.89, and 0.91, respectively. Compared with the BPNN, RBFNN, and RF models, the SVR model had the lowest errors, with mean MSE, RMSE, and MAPE values of 0.001, 0.02, and 2.77, respectively. The results show that the BPNN, RBFNN, RF, and SVR models are within acceptable ranges for karst NDVI prediction, but the overall performance of the SVR model is the best, and it is more suitable for karst vegetation prediction. [ABSTRACT FROM AUTHOR]

Coffee has high relevance in the Brazilian agricultural scenario, as Brazil is the largest producer and exporter of coffee in the world. Strategies to advance the production of coffee grains involve better understanding its spatial variability along fields. The objectives of this study were to adjust yield-prediction models based on a time series of satellite images and high-density yield data, and to indicate the best phenological stage of coffee crop to obtain satellite images for this purpose. The study was conducted during three seasons (2019, 2020 and 2021) in a commercial area (10.24 ha), located in the state of Minas Gerais, Brazil. Data were obtained using a harvester equipped with a yield monitor that measures the volume of coffee harvested with 3.0 m of spatial resolution. Satellite images from the PlanetScope (PS) platform were used. Random forest (RF) regression and multiple linear regression (MLR) models were fitted to different datasets composed of coffee yield and time series of satellite-image data ((1) Spectral bands—red, green, blue and near-infrared; (2) Normalized difference vegetation index (NDVI); or (3) Green normalized difference vegetation index (GNDVI)). Whether using RF or MLR, the spectral bands, NDVI and GNDVI reproduced the spatial variability of yield maps one year before harvest. This information can be of critical importance for management decisions across the season. For yield quantification, the RF model using spectral bands showed the best results, reaching R2 of 0.93 for the validation set, and the lowest errors of prediction. The most appropriate phenological stage for satellite-image data acquisition was the dormancy phase, observed during the dry season months of July and August. These findings can help to monitor the spatial and temporal variability of the fields and guide management practices based on the premises of precision agriculture. [ABSTRACT FROM AUTHOR]

Crop yield estimation has gained importance due to its vital significance for policymakers and decision-makers in enacting schemes, ensuring food security, and assessing crop insurance losses due to biotic and abiotic stress. This research focused on rice yield estimation at the field level in the Karimnagar district of Telangana during 2021 and 2022 by employing the leaf area index (LAI) as the primary criterion for integrating remote sensing technology and crop simulation models. Using Sentinel-2 satellite data, the rice crop was mapped with the help of ground data and machine learning algorithms, attaining an accuracy of 93.04%. Crop management data for the DSSAT tool were collected during the field visits; the model results revealed a 0.80 correlation between observed and predicted yields. Due to its strong correlation with LAI (0.82), the normalized difference vegetation index (NDVI) was selected as the critical element for integration with the model. A spatial LAI map was generated using the linear equation developed between the NDVI and LAI. The relationship between LAI and yield was used to create a spatial yield map. The study's findings show that assimilating remote sensing data with crop models enhances the precision of rice yield prediction for insurance companies and policy- and decision-makers. [ABSTRACT FROM AUTHOR]

The variable fertilization rate (VFR) technique has demonstrated its ability to reduce nutrient losses by adapting the fertilizer dose to crop needs. However, transferring this technology to farms is not easy. This study aimed to make a variable fertilization map in a commercial plot where there is no data from a yield monitor, combining machine learning techniques and farmer's knowledge. In addition to the normalized difference vegetation index (NDVI) obtained from Sentinel-2 and a digital elevation model (DEM), information captured by a yield monitor in 2019 was used to train and validate models. Among the 15 algorithms trained, the best result was obtained by the random forest (RF), with an RMSE of 496 and R2 of 0.90. Using the "leave one out" technique, the capacity to predict an entire plot was tested. Finally, the RF algorithm was tested on a 12-hectare wheat plot where no yield data were available. The novelty of this work lies in the collaborative work developed between farmers and researchers to implement the VRF technique in plots where precise yield data do not exist and in the "leave one out" validation. The collaboration between scientists and farmers resulted in a very positive exchange of information that allowed the farmer to change the fertilization strategy of the whole farm and the scientists to better understand how soil properties and plot history affect yield. [ABSTRACT FROM AUTHOR]

An accurate and universal monitoring model of nitrogen nutrition can greatly contribute to the precise management and application of regional rice. In this study, a systematic investigation was carried out to explore the impact of the images on the monitoring models. The consumer-grade multispectral images were captured by unmanned aerial vehicles (UAV) from the different ecological sites and rice varieties. A series of field experiments were conducted at two test sites: Mengzhe Town, Xishuangbanna, Yunnan Province (with trial variety Yunjing 37) and Beibei District, Chongqing City (with trial variety Jiyu 6135) in South China, with the varying nitrogen levels. Four multispectral drones (DJI Phantom) were used to capture the multispectral images of the rice canopy during tillering, jointing, and heading stages. The nitrogen content in rice plant canopies (CNC) was measured to calculate the above-ground nitrogen accumulation (PNA) using the Kjeldahl method. The nitrogen nutrition monitoring models were established using vegetation indices, partial least-squares regression (PLSR), random forest (RF), and backpropagation neural network (BPNN) for the single trial site, single variety, different trial sites, and multiple varieties of rice. The transferability was also explored in the models. The monitoring models were established for the CNC and PNA during each growth stage of rice for the single trial site and single variety with the high accuracy (normalized difference vegetation index NDVI or near-infrared normalized vegetation index NNVI, where the coefficient of determination was 0.68-0.88). However, the inaccurate model was obtained with the vegetation index during the tillering stage (NDVI, the coefficient of determination was 0.53-0.79), indicating the low transferability of all models. It was difficult for the vegetation indices to construct the monitoring models for the canopy nitrogen content throughout the entire growth period of rice. But a monitoring model was established for the above-ground nitrogen accumulation of single-variety rice in the period of the growth using the ratio vegetation index, the high accuracy and the high transferbility of the model. The monitoring models constructed by PLSR, RF, and BPNN were more accurate than those by vegetation indices. Among them, the highest accuracy was achieved in the monitoring model for the canopy nitrogen content and above-ground nitrogen accumulation in the entire growth period of multiple varieties using RF, with the coefficient of determination values were 0.84 and 0.94, respectively, and root-mean-square errors of 0.28% and 10.09 kg/hm², respectively. The finding can provide a theoretical basis and technical support for the application of consumer-grade multispectral images captured by drones in monitoring rice nitrogen nutrition during the entire growth period of different ecological sites and rice varieties. [ABSTRACT FROM AUTHOR]

The thickness or depth of fine-textured soil (z f) dominates water storage capacity and exerts a control on nutrient leaching in semi-arid agroecosystems. At small pixel sizes (< 1 m; 'fine resolution'), the normalized difference vegetation index (NDVI) of cereal crops during senescence (Zadoks Growth Stages [ZGS] 90–93) offers a promising alternative to destructive sampling of z f using soil pits. However, it is unclear whether correlations between z f and NDVI exist (a) at larger pixel sizes (1–10 m; 'intermediate resolution') and (b) across field boundaries. The relationship of z f to NDVI of wheat (Triticum aestivum L.) was tested using images from a combination of multispectral sensors and fields in central Montana. NDVI was derived for one field using sensors of fine and intermediate spatial resolution and for three fields using intermediate resolution sensors only. Among images acquired during crop senescence, z f was correlated with NDVI (p < 0.05) independent of sensor (p = 0.22) and field (p = 0.94). The z f relationship to NDVI was highly dependent on acquisition day (p < 0.05), but only when pre-senescence (ZGS ≤ 89) images were included in the analysis. Results indicate that cereal crop NDVI of intermediate resolution can be used to characterize z f across field boundaries if image acquisition occurs during crop senescence. Based on these findings, an empirical index was derived from multi-temporal Sentinel-2 imagery to estimate z f on fields in and beyond the study area. • Fine-textured soil thickness (z f) exerts a control on water storage and nitrate leaching. • z f was related to NDVI using images from a combination of sensors and dates. • NDVI of wheat during senescence was correlated with z f across field boundaries. • z f -NDVI relationships depended on acquisition date at early crop stages. • An empirical z f index was derived to improve precision water and nitrogen management. [ABSTRACT FROM AUTHOR]

Artisanal small-scale mines (ASMs) in the Amazon Rainforest are an important cause of deforestation, forest degradation, biodiversity loss, sedimentation in rivers, and mercury emissions. Satellite image data are widely used in environmental decision-making to monitor changes in the land surface, but ASMs are difficult to map from space. ASMs are small, irregularly shaped, unevenly distributed, and confused (spectrally) with other land clearance types. To address this issue, we developed a reliable and efficient ASM detection method for the Tapajós River Basin of Brazil—an important gold mining region of the Amazon Rainforest. We enhanced detection in three key ways. First, we used the time-series segmentation (LandTrendr) Google Earth Engine (GEE) Application Programming Interface to map the pixel-wise trajectory of natural vegetation disturbance and recovery on an annual basis with a 2000 to 2019 Landsat image time series. Second, we segmented 26 textural features in addition to 5 spectral features to account for the high spatial heterogeneity in ASM pixels. Third, we trained and tested a Random Forest model to detect ASMs after eliminating irrelevant and redundant features with the Variable Selection Using Random Forests "ensemble of ensembles" technique. The out-of-bag error and overall accuracy of the final Random Forest was 3.73 and 92.6%, which are comparable to studies mapping large industrial mines with the normalized difference vegetation index (NDVI) and LandTrendr. The most important feature in our study was NDVI, followed by textural features in the near and shortwave infrared. Our work paves the way for future ASM regulation through large area monitoring from space with free and open-source GEE and operational satellites. Studies with sufficient computational resources can improve ASM monitoring with advanced sensors consisting of spectral narrow bands (Sentinel-2, Environmental Mapping and Analysis Program, PRecursore IperSpettrale della Missione Applicativa) and deep learning. [ABSTRACT FROM AUTHOR]

Late leaf spot (LLS) is an important disease of peanut, causing global yield losses. Developing resistant varieties through breeding is crucial for yield stability, especially for smallholder farmers. However, traditional phenotyping methods used for resistance selection are laborious and subjective. Remote sensing offers an accurate, objective, and efficient alternative for phenotyping for resistance. The objectives of this study were to compare between regression and classification for breeding, and to identify the best models and indices to be used for selection. We evaluated 223 genotypes in three environments: Serere in 2020, and Nakabango and Nyankpala in 2021. Phenotypic data were collected using visual scores and two handheld sensors: a red–green–blue (RGB) camera and GreenSeeker. RGB indices derived from the images, along with the normalized difference vegetation index (NDVI), were used to model LLS resistance using statistical and machine learning methods. Both regression and classification methods were also evaluated for selection. Random Forest (RF), the artificial neural network (ANN), and k-nearest neighbors (KNNs) were the top-performing algorithms for both regression and classification. The ANN (R2: 0.81, RMSE: 22%) was the best regression algorithm, while the RF was the best classification algorithm for both binary (90%) and multiclass (78% and 73% accuracy) classification. The classification accuracy of the models decreased with the increase in classification classes. NDVI, crop senescence index (CSI), hue, and greenness index were strongly associated with LLS and useful for selection. Our study demonstrates that the integration of remote sensing and machine learning can enhance selection for LLS-resistant genotypes, aiding plant breeders in managing large populations effectively. [ABSTRACT FROM AUTHOR]

The application of remote sensing, which is non-destructive and cost-efficient, has been widely used in crop monitoring and management. This study used a built-in multispectral imager on a small unmanned aerial vehicle (UAV) to capture multispectral images in five different spectral bands (blue, green, red, red edge, and near-infrared), instead of satellite-captured data, to monitor soybean growth in a field. The field experiment was conducted in a soybean field at the Mississippi State University Experiment Station near Pontotoc, MS, USA. The experiment consisted of five cover crops (Cereal Rye, Vetch, Wheat, Mustard plus Cereal Rye, and native vegetation) planted in the winter and three fertilizer treatments (Fertilizer, Poultry Liter, and None) applied before planting the soybean. During the soybean growing season in 2022, eight UAV imaging flyovers were conducted, spread across the growth season. UAV image-derived vegetation indices (VIs) coupled with machine learning (ML) models were computed for characterizing soybean growth at different stages across the season. The aim of this study focuses on monitoring soybean growth to predict yield, using 14 VIs including CC (Canopy Cover), NDVI (Normalized Difference Vegetation Index), GNDVI (Green Normalized Difference Vegetation Index), EVI2 (Enhanced Vegetation Index 2), and others. Different machine learning algorithms including Linear Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) are used for this purpose. The stage of the initial pod development was shown as having the best predictability for earliest soybean yield prediction. CC, NDVI, and NAVI (Normalized area vegetation index) were shown as the best VIs for yield prediction. The RMSE was found to be about 134.5 to 511.11 kg ha−1 in the different yield models, whereas it was 605.26 to 685.96 kg ha−1 in the cross-validated models. Due to the limited number of training and testing samples in the K-fold cross-validation, the models' results changed to some extent. Nevertheless, the results of this study will be useful for the application of UAV remote sensing to provide information for soybean production and management. This study demonstrates that VIs coupled with ML models can be used in multistage soybean yield prediction at a farm scale, even with a limited number of training samples. [ABSTRACT FROM AUTHOR]

While protected areas (PAs) are an important conservation strategy to protect vulnerable ecosystems and species, recent analyses question their effectiveness in curbing deforestation and maintaining landscape connectivity. The spatial arrangement of forests inside and outside of PAs may affect ecosystem functioning and wildlife movement. The Osa Peninsula—and Costa Rica in general—are unique conservation case studies due to their high biodiversity, extensive PA network, environmental policies, and payment for ecosystem services (PES) programs. This study explores the relationship between forest management initiatives—specifically PAs, the 1996 Forest Law, and PES—and forest cover and landscape metrics in the Osa Conservation Area (ACOSA). The Google Earth Engine API was used to process Surface Reflectance Tier 1 Landsat 5 Thematic Mapper and Landsat 8 Operational Land Imager data for 1987, 1998, and 2019, years with relatively cloud-free satellite imagery. Land use/land cover (LULC) maps were generated with the pixel-based random forest machine learning algorithm, and Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and functional landscape metrics were calculated. The LULC maps are the first to track land use change, from 1987 to 2019 and the first to separately classify mature and secondary forest in the region, and have already proven useful for conservation efforts. The results suggest that forest cover, NDVI, EVI, and structural connectivity increased from 1987 to 2019 across the study area, both within and surrounding the PAs, suggesting minimal deforestation encroachment and local leakage. These changes may have contributed to the increasing vertebrate abundance observed in the region. PAs, especially national parks with stricter conservation regulations, displayed the highest forest cover and connectivity. Forest cover increased in properties receiving PES payments. Following the Forest Law's 1996 deforestation ban, both forest conversion and reforestation rates decreased, suggesting the law curbed deforestation but did not drive reforestation across the region. Connectivity outside of PAs slightly declined following the adoption of the law, so the subsequent forest growth likely occurred in mostly previously unforested areas. Forest expansion alone does not ensure connectivity. We highlight the importance of developing policies, PES programs, and monitoring systems that emphasize conserving and restoring large, connected forest patches for biodiversity conservation and landscape resilience. Resumen: Aunque las áreas protegidas (APs) son una importante estrategia de conservación para proteger ecosistemas y especies vulnerables, algunos análisis recientes cuestionan su eficacia para frenar la deforestación y mantener la conectividad del paisaje. La distribución espacial de los bosques dentro y fuera de las AP puede afectar el funcionamiento de los ecosistemas y los movimientos de la fauna. La Península de Osa–y Costa Rica en general–constituyen casos de estudio únicos de conservación debido a su elevada biodiversidad, su extensa red de AP, sus políticas medioambientales y sus programas de Pago por Servicios Ambientales (PSA). Este estudio explora la relación entre APs, la Ley Forestal de 1996, PSA, cobertura y métricas del paisaje en el Área de Conservación Osa (ACOSA). Se utilizó la plataforma Google Earth Engine API para procesar datos de Reflectancia Superficial Tier 1 Landsat 5 Thematic Mapper y Landsat 8 Operational Land Imager para 1987, 1998 y 2019, años con imágenes satelitales relativamente libres de nubes. Se generaron mapas de uso del suelo con el algoritmo de aprendizaje automático basado en pixeles Random Forest, y se calcularon el índice de vegetación de diferencia normalizada (NDVI), el índice de vegetación mejorado (EVI) y las métricas de paisaje funcionales. Estos mapas, los primeros en clasificar por separado los bosques maduros y secundarios de la región, han demostrado su utilidad para los esfuerzos de conservación. Los resultados sugieren que la cobertura forestal, el NDVI, el EVI y la conectividad estructural aumentaron entre 1987 y 2019 en toda la región de estudio, tanto dentro de las AP como en sus alrededores, lo que sugiere una expansión mínima de la deforestación dentro y fuera de las AP. Estos cambios pueden haber contribuido al aumento de la abundancia de vertebrados observado en la región. Las AP, especialmente los parques nacionales con regulaciones de conservación más estrictas, mostraron la mayor cobertura forestal y conectividad. La cobertura forestal aumentó en aquellas propiedades que recibieron PSA. Tras la prohibición de la deforestación por la Ley Forestal de 1996, disminuyeron tanto las tasas de conversión forestal como las de reforestación, lo que sugiere que la ley frenó la deforestación, pero no impulsó la reforestación. La conectividad fuera de las AP disminuyó ligeramente tras la entrada en vigor de la ley, lo que sugiere que el crecimiento forestal posterior se produjo en zonas que antes no estaban forestadas. Por lo tanto, la expansión forestal por sí sola no garantiza la conectividad. Resaltamos la importancia de desarrollar políticas, programas PSA y sistemas de monitoreo que hagan hincapié en la conservación y restauración de grandes zonas forestales conectadas para apuntalar la conservación de la biodiversidad y la resiliencia del paisaje. [ABSTRACT FROM AUTHOR]

Irrigation stands as the primary sector of human water consumption and plays a pivotal role in enhancing crop yields and mitigating drought effects. The precise distribution of irrigation is crucial for effective water resource management and the assessment of food security. However, the existing global irrigated cropland map is characterized by a coarse resolution, typically around 10 kilometers, and is often not regularly updated. In our study, we present a robust methodology that leverages irrigation performance during drought stress as an indicator of crop productivity and water consumption to identify global irrigated cropland. Within each irrigation mapping zone (IMZ), we identified the dry months occurring during the growing season from 2017 to 2019 or the driest months from 2010 to 2019. To delineate irrigated cropland, we utilized collected samples to calculate normalized difference vegetation index (NDVI) thresholds for the dry months of 2017 to 2019 and the NDVI deviation from the ten-year average for the driest month. By combining the results with the higher accuracy between these two methods, we generated the Global Maximum Irrigation Extent dataset at 100-meter resolution (GMIE-100), achieving an overall accuracy of 83.6%. GMIE-100 reveals that the maximum extent of irrigated cropland encompasses 403.17 million hectares, accounting for 23.4% of the global cropland. Concentrated in fertile plains and regions adjacent to major rivers, the largest irrigated cropland area is found in India, followed by China, the United States, and Pakistan, ranking 2nd to 4th, respectively. Importantly, the spatial resolution of GMIE-100, at 100 meters, surpasses that of the dominant irrigation map, offering more detailed information essential for supporting estimates of agricultural water use and regional food security assessments. Furthermore, with the help of the DL method, the global central pivot irrigation system (CPIS) was identified using Pivot-Net. We found that there are 11.5 million hectares of CPIS, accounting for about 2.9% of total irrigated cropland. In Namibia, the US, Saudi Arabia, South Africa, Canada, and Zambia, the CPIS proportion was larger than 10%. To our best knowledge, this is the first effort to identify irrigation methods globally. The GMIE-100 dataset containing both or irrigated extent and CPIS distribution is accessible on Harvard Dataverse at: https://doi.org/10.7910/DVN/HKBAQQ (Tian et al., 2023a). [ABSTRACT FROM AUTHOR]

The Global Navigation Satellite System Interferometric Reflectometry (GNSS-IR) technique provides a new remote sensing method that shows great potential for soil moisture detection and vegetation growth, as well as for climate research, water cycle management, and ecological environment monitoring. Considering that the land surface is always covered by vegetation, it is essential to take into account the impacts of vegetation growth when detecting soil moisture (SM). In this paper, based on the GNSS-IR technique, the SM was retrieved from multi-GNSS and multi-frequency data using a machine learning model, accounting for the impact of the vegetation moisture content (VMC). Both the signal-to-noise ratio (SNR) data that was used to retrieve SM and the multipath data that was used to eliminate the vegetation influence were collected from a standard geodetic GNSS station located in Nanjing, China. The normalized microwave reflectance index (NMRI) calculated by multipath data was mapped to a normalized difference vegetation index (NDVI), which was derived from Sentinel-2 data on the Google Earth Engine platform to estimate and eliminate the influence of VMC. Based on the characteristic parameters of amplitude and phase extracted from detrended SNR signals and NDVI derived from multipath data, three machine learning methods, including random forest (RF), multiple linear regression (MLR), and multivariate adaptive regression spline (MARS), were employed for data fusion. The results show that the vegetation effect can be well eliminated using the NMRI method. Comparing MLR and MARS, RF is more suitable for GNSS-IR SM inversion. Furthermore, the SM reversed from amplitude and phase fusion is better than only those from either amplitude fusion or phase fusion. The results prove the feasibility of the proposed method based on a multipath approach to characterize the vegetation effect, as well as the RF model to fuse multi-GNSS and multi-frequency data to retrieve SM with vegetation error-correcting. [ABSTRACT FROM AUTHOR]

Vegetation is a key indicator of the health of most terrestrial ecosystems and different types of vegetation exhibit different sensitivity to climate change. The Yarlung Zangbo River Basin (YZRB) is one of the highest basins in the world and has a wide variety of vegetation types because of its complex topographic and climatic conditions. In this paper, the sensitivity to climate change for different vegetation types, as reflected by the Normalized Difference Vegetation Index (NDVI), was assessed in the YZRB. Three machine learning models, including multiple linear regression, support vector machine, and random forest, were adopted to simulate the response of each vegetation type to climatic variables. We selected random forest, which showed the highest performance in both the calibration and validation periods, to assess the sensitivity of the NDVI to temperature and precipitation changes on an annual and monthly scale using hypothetical climatic scenarios. The results indicated there were positive responses of the NDVI to temperature and precipitation changes, and the NDVI was more sensitive to temperature than to precipitation on an annual scale. The NDVI was predicted to increase by 1.60–4.68% when the temperature increased by 1.5 °C, while it only changed by 0.06–0.24% when the precipitation increased by 10% in the YZRB. Monthly, the vegetation was more sensitive to temperature changes in spring and summer. Spatially, the vegetation was more sensitive to temperature increases in the upper and middle reaches, where the existing temperatures were cooler. The time-lag effects of climate were also analyzed in detail. For both temperature and precipitation, Needleleaf Forest and Broadleaf Forest had longer time lags than those of other vegetation types. These findings are useful for understanding the eco-hydrological processes of the Tibetan Plateau. [ABSTRACT FROM AUTHOR]

In dry regions, gardens and trees within the urban space are of considerable significance. These gardens are facing harsh weather conditions and environmental stresses; on the other hand, due to the high value of land in urban areas, they are constantly subject to destruction and land use change. Therefore, the identification and monitoring of gardens in urban areas in dry regions and their impact on the ecosystem are the aims of this study. The data utilized are aerial and Sentinel-2 images (2018–2022) for Yazd Township in Iran. Several satellite and aerial image fusion methods were employed and compared. The root mean square error (RMSE) of horizontal shortcut connections (HSC) and color normalization (CN) were the highest compared to other methods with values of 18.37 and 17.5, respectively, while the Ehlers method showed the highest accuracy with a RMSE value of 12.3. The normalized difference vegetation index (NDVI) was then calculated using the images with 15 cm spatial resolution retrieved from the fusion. Aerial images were classified by NDVI and digital surface model (DSM) using object-oriented methods. Different object-oriented classification methods were investigated, including support vector machine (SVM), Bayes, random forest (RF), and k-nearest neighbor (KNN). SVM showed the greatest accuracy with overall accuracy (OA) and kappa of 86.2 and 0.89, respectively, followed by RF with OA and kappa of 83.1 and 0.87, respectively. Separating the gardens using NDVI, DSM, and aerial images from 2018, the images were fused in 2022, and the current status of the gardens and associated changes were classified into completely dried, drying, acceptable, and desirable conditions. It was found that gardens with a small area were more prone to destruction, and 120 buildings were built in the existing gardens in the region during 2018–2022. Moreover, the monitoring of land surface temperature (LST) showed an increase of 14 °C in the areas that were changed from gardens to buildings. [ABSTRACT FROM AUTHOR]

The leaf area index (LAI), a crucial biophysical indicator, is used to assess and monitor crop growth for effective agricultural management. This study assessed the LAI at the seedling stage after conducting a field experiment with rainfed groundnut. The study tests the performance of multiple machine learning regression algorithms (MLRAs) and empirical vegetation indices (VIs) in retrieving groundnut's LAI using freely available Sentinel-2 data. The bands at 665 nm, 705 nm, 842 nm, and 2190 nm are the most sensitive for retrieving groundnut's LAI, according to an analysis of its band spectrum. Results suggest that VIs computed with wavebands centered at red (665 nm), red edge (705 nm), and near-infrared (842 nm) exhibited optimal R2 with Sentinel-2 data. Normalized difference vegetation index (NDVI), red edge normalized difference vegetation index (NDVIre), simple ratio (SR), red edge simple ratio (SRre), and green normalized difference vegetation index (gNDVI) were utilized as predictors for LAI. Regarding the results of the validation between estimated and measured LAI, SR demonstrated the highest accuracy for groundnut LAI prediction (r2 = 0.67, RMSE = 0.89). Ten MLRAs were tested, and results indicate from the perspective of the accuracy of models, the Gaussian processes regression, GPR (r2 = 0.73 and RMSE = 0.81), Kernel ridge regression, KRR (r2 = 0.72 and RMSE = 0.82) and Support vector regression, SVR (r2 = 0.70 and RMSE = 0.85) demonstrated to be the most suitable for LAI estimation for rainfed groundnut at the seedling stage. The systematic analysis based on the regression approaches tested here revealed that the GPR outperformed other models combined, therefore, most suitable for estimating rainfed groundnut LAI at the seedling stage. These findings serve as a benchmark for obtaining crop biophysical parameters in the framework of groundnut traits monitoring in a tropical West Africa. [ABSTRACT FROM AUTHOR]

Introduction-- Drought is one of the most critical hydrometeorological phenomenon in terms of its impacts on society. Although Colombia is a tropical country, there are areas of the territory which have periods of drought, and this causes significant economic damage. Objective-- Due to recent advances in terms of the spatial and temporal resolutions of remote sensing, and artificial intelligence techniques, it is possible to develop automatic learning models supported by historical information. Methodology-- In this study, a Random Forest (RF) and Bagged Decision Tree Classifier (DTC) model was built to perform spatial and temporal drought prediction in the department of Magdalena using the following features: Normalized Difference Vegetation Index (NDVI), land surface temperature (LST), precipitation, Normalized Difference Water Index (NDWI), Normalized Multiband Drought Index (NMDI), evapotranspiration (ET), surface soil moisture (SSM), subsurface soil moisture (SUSM), Multivariate ENSO Index (MEI), Southern Oscillation Index (SOI), and Oceanic Niño Index (ONI). Results-- For labelling, which allows one to train and evaluate the model, the Standardized Precipitation Index (SPI) was used to identify drought events. Conclusions-- The implementation of the developed model can allow governmental entities to take actions to mitigate impacts generated by recurring droughts in their territories. [ABSTRACT FROM AUTHOR]

Agricultural intensification practices have been adopted in the Brazilian savanna (Cerrado), mainly in the transition between Cerrado and the Amazon Forest, to increase productivity while reducing pressure for new land clearing. Due to the growing demand for more sustainable practices, more accurate information on geospatial monitoring is required. Remote sensing products and artificial intelligence models for pixel-by-pixel classification have great potential. Therefore, we developed a methodological framework with spectral indices (Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), and Soil-Adjusted Vegetation Index (SAVI)) derived from the Harmonized Landsat Sentinel-2 (HLS) and machine learning algorithms (Random Forest (RF), Artificial Neural Networks (ANNs), and Extreme Gradient Boosting (XGBoost)) to map agricultural intensification considering three hierarchical levels, i.e., temporary crops (level 1), the number of crop cycles (level 2), and the crop types from the second season in double-crop systems (level 3) in the 2021–2022 crop growing season in the municipality of Sorriso, Mato Grosso State, Brazil. All models were statistically similar, with an overall accuracy between 85 and 99%. The NDVI was the most suitable index for discriminating cultures at all hierarchical levels. The RF-NDVI combination mapped best at level 1, while at levels 2 and 3, the best model was XGBoost-NDVI. Our results indicate the great potential of combining HLS data and machine learning to provide accurate geospatial information for decision-makers in monitoring agricultural intensification, with an aim toward the sustainable development of agriculture. [ABSTRACT FROM AUTHOR]

Plants respond to biotic and abiotic pressures by changing their biophysical and biochemical aspects, such as reducing their biomass and developing chlorosis, which can be readily identified using remote-sensing techniques applied to the VIS/NIR/SWIR spectrum range. In the current scenario of agriculture, production efficiency is fundamental for farmers, but diseases such as target spot continue to harm soybean yield. Remote sensing, especially hyperspectral sensing, can detect these diseases, but has disadvantages such as cost and complexity, thus favoring the use of UAVs in these activities, as they are more economical. The objectives of this study were: (i) to identify the most appropriate input variable (bands, vegetation indices and all reflectance ranges) for the metrics assessed in machine learning models; (ii) to verify whether there is a statistical difference in the response of NDVI (normalized difference vegetation index), grain weight and yield when subjected to different levels of severity; and (iii) to identify whether there is a relationship between the spectral bands and vegetation indices with the levels of target spot severity, grain weight and yield. The field experiment was carried out in the 2022/23 crop season and involved different fungicide treatments to obtain different levels of disease severity. A spectroradiometer and UAV (unmanned aerial vehicle) imagery were used to collect spectral data from the leaves. Data were subjected to machine learning analysis using different algorithms. LR (logistic regression) and SVM (support vector machine) algorithms performed better in classifying target spot severity levels when spectral data were used. Multivariate canonical analysis showed that healthy leaves stood out at specific wavelengths, while diseased leaves showed different spectral patterns. Disease detection using hyperspectral sensors enabled detailed information acquisition. Our findings reveal that remote sensing, especially using hyperspectral sensors and machine learning techniques, can be effective in the early detection and monitoring of target spot in the soybean crop, enabling fast decision-making for the control and prevention of yield losses. [ABSTRACT FROM AUTHOR]

Precision viticulture systems are essential for enhancing traditional intensive viticulture, achieving high-quality results, and minimizing costs. This study explores the integration of Unmanned Aerial Vehicles (UAVs) and artificial intelligence in precision viticulture, focusing on vine detection and vineyard zoning. Vine detection employs the YOLO (You Only Look Once) deep learning algorithm, achieving a remarkable 90% accuracy by analysing UAV imagery with various spectral ranges from various phenological stages. Vineyard zoning, achieved through the application of the K-means algorithm, incorporates geospatial data such as the Normalized Difference Vegetation Index (NDVI) and the assessment of nitrogen, phosphorus, and potassium content in leaf blades and petioles. This approach enables efficient resource management tailored to each zone's specific needs. The research aims to develop a decision-support model for precision viticulture. The proposed model demonstrates a high vine detection accuracy and defines management zones with variable weighting factors assigned to each variable while preserving location information, revealing significant differences in variables. The model's advantages lie in its rapid results and minimal data requirements, offering profound insights into the benefits of UAV application for precise vineyard management. This approach has the potential to expedite decision making, allowing for adaptive strategies based on the unique conditions of each zone. [ABSTRACT FROM AUTHOR]

The fruit supply chain (FSC) involves different stages that need to be planned at least two months in advance. Therefore, having a good fruit yield forecast with anticipation allows making timely correct decisions for providing the resources, transport, and cold storage contracts, among others. Therefore, fruit yield over or underestimation could cause important inefficiencies with regards to FSC. Because of its relevance, a method based on machine learning (ML) techniques that uses spectrometric vegetation data is proposed. This method, known as Spectrometry Based Method for Fruit Production Forecast (SBM-Fruit), allows exploring the georeferenced Normalized Difference Vegetation Index (NDVI), collected in different phenological stages, aiming to capture spatial and temporal dependency in the fruit yield forecast. In the first step of SBM-Fruit, several clusters are obtained in a clustering process using the georeferenced NDVI in all phenological stages as input, while, in the second step, two validation functions are used for determining the best clustering. Finally, in the third step, the predictor variables of the best clustering are incorporated into an artificial neural network (ANN) for predicting the fruit yield. The SBM-Fruit was applied to forecast table grape yield of an orchard located in the Valparaíso Region, Chile. The results show fruit yield estimations with mean errors around 0.013 percent for every spatial zone of the orchard, forecasted at least two months in advance. The use of the SBM-Fruit would allow FSC stakeholders to make better decisions, improving the coordination of the FSC stages, and reducing costs and fruit losses. [ABSTRACT FROM AUTHOR]

Natural resources are under tremendous amounts of threat as a result of the expanding human population, which over time intensifies changes in Land use and Land cover (LULC). Understanding how various machine learning classifiers function is crucial as the demand for an accurate estimate of LULC from satellite images. The purpose of this research was to classify the LULC in the entire Karnataka state, using three distinct methods on the Google Earth Engine (GEE) namely RF (Random Forest), SVM (Support Vector Machine) and CART (Classification Regression Trees), are examples of machine learning techniques. The LULC is classified by the training sets using supervised classification. The NDVI (Normalized difference vegetation index) was assessed and used to increase classification accuracy. The LULC classification for the years 2015 to 2021 utilizes multi-temporal images like Sentinel-2, Landsat-8, and MODIS data with spatial resolution of 10 m, 30 m, and 250 m. Agricultural land, Built-up land, Forest land, Fallow land, wasteland, water body and others, are major LULC classes, it lies on a level I classification. According to the findings, the change % of agricultural land is high from 2015 (64.03%) to 2021 (67.81%), this roughly increased about 3.78% during the study year. While water bodies increased by 5.25 to 6.3%. Based on the results, the largest LULC group is agricultural land (122,789.4 km2 or 64.03%), followed by forest land (37,678.56 km2 or 19.65%). Increased built-up land in the studied area indicates extraordinarily rapid urban growth in major cities of the state. This research offers a reliable approach for comprehensive, automated, and LULC classification in Karnataka State. [ABSTRACT FROM AUTHOR]

Large wildfires (>125 hectares) in the United States account for over 95% of the burned area each year. Predicting large wildfires is imperative; however, current wildfire predictive models are region-based and computationally intensive. Using a scalable model based on easily available environmental and atmospheric data, this research aims to accurately predict whether large wildfires will develop across the United States. The data used in this study include 2109 wildfires over 20 years, representing 14 million hectares burned. Remote sensing environmental data (Normalized Difference Vegetation Index—NDVI; Enhanced Vegetation Index—EVI; Leaf Area Index—LAI; Fraction of Photosynthetically Active Radiation—FPAR; Land Surface Temperature during the Day—LST Day; and Land Surface Temperature during the Night—LST Night) consisting of 1.3 billion satellite observations was used. Atmospheric reanalysis data (u component of wind, v component of wind, relative humidity, temperature, and geopotential) at four pressure levels (300, 500, 700, and 850 Ha) were also factored in. Six machine learning classification models (Logistic Regression, Decision Tree, Random Forest, eXtreme Gradient Boosting, K-Nearest Neighbors, and Support Vector Machine) were created and tested on the resulting dataset to determine their accuracy in predicting large wildfires. Model validation tests and variable importance analysis were performed. The eXtreme Gradient Boosting (XGBoost) classification model performed best in predicting large wildfires, with 90.44% accuracy, a true positive rate of 0.92, and a true negative rate of 0.88. Furthermore, towards environmental justice, an analysis was performed to identify disadvantaged communities that are also vulnerable to wildfires. This model can be used by wildfire safety organizations to predict large wildfires with high accuracy and prioritize resource allocation to employ protective safeguards for impacted socioeconomically disadvantaged communities. [ABSTRACT FROM AUTHOR]

Precise and cost-effective mapping of crop and other land cover (LC) types is essential for food security, precision agriculture, and water distribution management. However, accurate identification of the extent of crop and LC types is still a challenge. However, the efficiency of machine learning (ML) and deep learning (DL) algorithms were evaluated and assimilated with time-series satellite information and In-situ data for detecting crop and other LC types. Therefore, in the current study, a framework was established on a deep convolutional neural network (CNN) using Sentinel-2 time-series satellite datasets to identify rice crop and other LC types without survey data. The normalized difference vegetation index (NDVI) stack was developed through Google Earth Engine (GEE) and adopted as input features. Three widely used ML algorithms, random forest (RF), support vector machine (SVM) and classification and regression trees (CART), and four DL models (Swin Transformer (ST), HRNet, two-dimensional convolutional neural network (2D-CNN) and long short-term memory (LSTM) were employed to perform the classification of rice crop and four non-crop classes. The overall accuracies were determined to be 93.95%, 91.67%, 89.80%, 86.89%, 81.01%, 76.51%, and 72.9% and the kappa coefficients were 92.35%, 89.49%, 87.13%, 83.52%, 76.12%, 70.56%, and 66.56%, of corresponding methods respectively. The findings showed that DL methods outperformed ML methods, while ST methods yielded the highest accuracy, and input data can be prepared for LULC classification without ground survey. Accurate classification of LC types and timely assessment of rice crop estimation can provide valuable statistics for state officials, decision-makers, and planners. [ABSTRACT FROM AUTHOR]

Infection of liver flukes (Opisthorchis viverrini) is partly due to their suitability for habitats in sub-basin areas, which causes the intermediate host to remain in the watershed system in all seasons. The spatial monitoring of fluke at the small basin scale is important because this can enable analysis at the level of the factors involved that influence infections. A spatial mathematical model was weighted by the nine spatial factors X1 (index of land-use types), X2 (index of soil drainage properties), X3 (distance index from the road network, X4 (distance index from surface water resources), X5 (distance index from the flow accumulation lines), X6 (index of average surface temperature), X7 (average surface moisture index), X8 (average normalized difference vegetation index), and X9 (average soil-adjusted vegetation index) by dividing the analysis into two steps: (1) the sub-basin boundary level was analyzed with an ordinary least square (OLS) model used to select the spatial criteria of liver flukes aimed at analyzing the factors related to human liver fluke infection according to sub-watersheds, and (2) we used the infection risk positional analysis level through machine-learning-based forest classification and regression (FCR) to display the predictive results of infection risk locations along stream lines. The analysis results show four prototype models that import different independent variable factors. The results show that Model 1 and Model 2 gave the most AUC (0.964), and the variables that influenced infection risk the most were the distance to stream lines and the distance to water bodies; the NDMI and NDVI factors rarely affected the accuracy. This FCR machine-learning application approach can be applied to the analysis of infection risk areas at the sub-basin level, but independent variables must be screened with a preliminary mathematical model weighted to the spatial units in order to obtain the most accurate predictions. [ABSTRACT FROM AUTHOR]

The traditional method for determining wine grape total soluble solid (TSS) is destructive laboratory analysis, which is time consuming and expensive. In this study, we explore the potential of using different predictor variables from various advanced techniques to predict the grape TSS in a non-destructive and rapid way. Calculating Pearson's correlation coefficient between the vegetation indices (VIs) obtained from UAV multispectral imagery and grape TSS resulted in a strong correlation between OSAVI and grape TSS with a coefficient of 0.64. Additionally, seven machine learning models including ridge regression and lasso regression, k-Nearest neighbor (KNN), support vector regression (SVR), random forest regression (RFR), extreme gradient boosting (XGBoost), and artificial neural network (ANN) are used to build the prediction models. The predictor variables include the unmanned aerial vehicles (UAV) derived VIs, and other ancillary variables including normalized difference vegetation index (NDVI_proximal) and soil electrical conductivity (ECa) measured by proximal sensors, elevation, slope, trunk circumference, and day of the year for each sampling date. When using 23 VIs and other ancillary variables as input variables, the results show that ensemble learning models (RFR, and XGBoost) outperform other regression models when predicting grape TSS, with the average of root mean square error (RMSE) of 1.19 and 1.2 °Brix, and coefficient of determination (R2) of 0.52 and 0.52, respectively, during the 20 times testing process. In addition, this study examines the prediction performance of using optimized soil adjusted vegetation index (OSAVI) or normalized green-blue difference index (NGBDI) as the main input for different machine learning models with other ancillary variables. When using OSAVI-based models, the best prediction model is RFR with an average R2 of 0.51 and RMSE of 1.19 °Brix, respectively. For NGBDI-based model, the RFR model showed the best average result of predicting TSS were a R2 of 0.54 and a RMSE of 1.16 °Brix, respectively. The approach proposed in this study provides an opportunity to grape growers to estimate the whole vineyard grape TSS in a non-destructive way. [ABSTRACT FROM AUTHOR]

Understanding the relationships between crop yields, soil properties, weather patterns and input applications is important for optimizing agricultural production. Data variation analysis using statistical and machine learning (ML) approaches can help identify and understand the practices that optimize yield. The objectives of this study were (i) to evaluate the predictive accuracy of selected ML models for estimating grain yields in on-farm maize (Zea mays L.) trials with different combinations of seeding and fertilizer rates in a commercial field, and (ii) to investigate the ability of ML models to assist in identifying yield-limiting factors in the same field. Multiple linear regression, multilayer perceptron, decision tree, and random forest (RF) ML models were trained and tested using crop management and soil from a data-intensive farm management (DIFM) trial and remotely sensed data. The dataset consisted of multiple subplot treatment observations of crop management, soil properties and normalized difference vegetation index (NDVI), linked to final grain yield for the 2019/2020 and 2020/2021 seasons. The RF had the best combination of high correlation (R2 = 0.69 and 0.80) and low error (MAPE = 5.4 and 8.4% and RMSE = 0.69 and 0.95 t ha−1) when compared to other models for both seasons. Feature importance analysis revealed that urea application was consistently the most critical variable and explained yield variations to the greatest extent, whereas soil phosphorus (P), plant population, and sodium in 2020, and soil P, soil pH, clay content, and plant population in 2021 emerged as the most influential factors for explaining yields. This study concluded that the RF model was the best for spatial yield predictions using DIFM trial datasets. There was also variability between seasons in yield limiting factors resulting from temporal variations in growing conditions. To effectively apply insights from yield prediction models, it is crucial that the variables incorporated into these models have a significant connection to yield and the findings can be translated into actionable management decisions. The DIFM trials combined with ML can play an important role in advancing the field of precision agriculture by providing valuable insights into the complex interactions between crops, soils, and management practices, and identifying new opportunities for improving crop yields and environmental sustainability. • Machine learning (ML) models were developed using Data Intensive Farm Management (DIFM) datasets. • The random forest was the best ML model for spatial yield prediction. • The extent at which individual soil and agronomic attributes limit crop yield can vary between seasons. • The inclusion of in-season NDVI data enhanced yield prediction accuracy of ML models. [ABSTRACT FROM AUTHOR]

• A bayesian neural network model predicts Canadian spring wheat yields. • Inputs to model includes NDVI anomalies and drought assessments. • The model predicts yield with errors ranging from 13% to 18% of mean yields. • Model can be used with drought assessments to understand impact of drought hazards on risks to crop yields. The Canadian drought monitor (CDM) provides monthly assessments of drought severity but does not provide guidance on the potential agricultural impacts. This study extends the CDM to a comprehensive risk assessment framework by assessing the vulnerability of spring wheat to drought timing and severity in the Canadian Prairies. To achieve this, a bayesian neural network (BNN) model was developed that predicts high-resolution wheat yields and uncertainties using the CDM and normalized difference vegetation index (NDVI) data from 2010 to 2021. Model performance is comparable to previous studies, with errors ranging from 13% to 18% of mean yields and R2 values from 0.43 to 0.70 that peak in early August. The model was then applied under various drought conditions to evaluate how different drought timings and severities affect yield vulnerability. The results demonstrate that drought exposure from June to August reduces wheat yields in proportion to drought severity while abnormally dry conditions may have a positive impact of yield, particularly in August. Yield uncertainty decreases substantially from June to July while yield prediction accuracy increases, suggesting that mitigation efforts are most effective prior to July. The final component of the research involved creating operational risk maps that integrate the yield vulnerability assessment with the CDM. These maps were applied in a case study to demonstrate their utility in providing actionable, real-time information for agricultural stakeholders. Overall, this study contributes to the body of knowledge regarding crop sensitivities to varying drought conditions and represents a methodological advancement by enabling high-resolution, operational agricultural risk assessments at scale. By integrating detailed vulnerability data with drought hazard monitoring from the CDM, the information provided by these risk maps can be leveraged by agricultural stakeholders in near real-time to make well-informed decisions that mitigate the impacts of drought on wheat production, enhancing the resilience of agricultural systems to climatic variability. [ABSTRACT FROM AUTHOR]

Tree species maps derived from satellite imagery increasingly support forest administrations and nature conservation authorities with large-scale and up-to-date information. However, many species are often excluded or aggregated in classification tasks due to a limited knowledge of the most suitable predictors. Our study aims to gain a better understanding of optical and polarimetric traits for tree species mapping by examining Sentinel-1 and Sentinel-2 time series from 61 tree species in temperate Europe. For a selection of 32 optical, polarimetric and structural variables, the principal component analysis revealed that Sentinel-2 variables mainly explain the variance in the data by contributing to the "seasonality" and "foliage color" components. Sentinel-1 contribute most to the "texture" component. The Normalized Difference Vegetation Index (NDVI), Tasseled Cap Greenness (TCG) and Radar Vegetation Index (RVI) were chosen as key variables for further analysis. Seasonality was found to be the most dominant aspect in all vegetation indices. Furthermore, the TCG was found to be useful to distinguish between early and late budding species. The RVI showed a large potential to discriminate conifers, which is attributed to the crown volume effect of C-band SAR. Using exploratory data analysis, we further examined the influence of management, biogeographical and meteorological factors on the time series from Fagus sylvatica , Pinus sylvestris , and Picea abies. The NDVI and TCG are relatively robust to different conditions. For the two conifer species however, we found strong spatial variations of the RVI which are presumably caused by different crown conditions across the study area. Using Sentinel-1 data could therefore lead to uncertainties in tree species mapping across large biogeographical gradients. This study contributes to the improvement of tree species mapping based on optical and dual-polarimetric data and thus benefits forest authorities and other stakeholders in their monitoring tasks and decision-making. [Display omitted] • The spectral-temporal characteristics of 61 tree species are investigated. • S1 and S2 mainly depict "seasonality", "foliage color" and "texture" information. • Grouping species by their spectral-temporal traits is better than using the genus. • Volume scattering at C-band SAR has a great potential to discriminate conifers. • S1 backscatter at spruce and pine stands is strongly affected by biogeography. [ABSTRACT FROM AUTHOR]

Improving high-resolution (meter-scale) mapping of snow-covered areas in complex and forested terrains is critical to understanding the responses of species and water systems to climate change. Commercial high-resolution imagery from Planet Labs, Inc. (Planet, San Francisco, CA, USA) can be used in environmental science, as it has both high spatial (0.7–3.0 m) and temporal (1–2 day) resolution. Deriving snow-covered areas from Planet imagery using traditional radiometric techniques have limitations due to the lack of a shortwave infrared band that is needed to fully exploit the difference in reflectance to discriminate between snow and clouds. However, recent work demonstrated that snow cover area (SCA) can be successfully mapped using only the PlanetScope 4-band (Red, Green, Blue and NIR) reflectance products and a machine learning (ML) approach based on convolutional neural networks (CNN). To evaluate how additional features improve the existing model performance, we: (1) build on previous work to augment a CNN model with additional input data including vegetation metrics (Normalized Difference Vegetation Index) and DEM-derived metrics (elevation, slope and aspect) to improve SCA mapping in forested and open terrain, (2) evaluate the model performance at two geographically diverse sites (Gunnison, Colorado, USA and Engadin, Switzerland), and (3) evaluate the model performance over different land-cover types. The best augmented model used the Normalized Difference Vegetation Index (NDVI) along with visible (red, green, and blue) and NIR bands, with an F-score of 0.89 (Gunnison) and 0.93 (Engadin) and was found to be 4% and 2% better than when using canopy height- and terrain-derived measures at Gunnison, respectively. The NDVI-based model improves not only upon the original band-only model's ability to detect snow in forests, but also across other various land-cover types (gaps and canopy edges). We examined the model's performance in forested areas using three forest canopy quantification metrics and found that augmented models can better identify snow in canopy edges and open areas but still underpredict snow cover under forest canopies. While the new features improve model performance over band-only options, the models still have challenges identifying the snow under trees in dense forests, with performance varying as a function of the geographic area. The improved high-resolution snow maps in forested environments can support studies involving climate change effects on mountain ecosystems and evaluations of hydrological impacts in snow-dominated river basins. [ABSTRACT FROM AUTHOR]

With the development of multisource satellite platforms and the deepening of remote sensing applications, the growing demand for high-spatial resolution and high-temporal resolution remote sensing images has aroused extensive interest in spatiotemporal fusion research. However, reducing the uncertainty of fusion results caused by sensor inconsistencies and input data preprocessing is one of the challenges in spatiotemporal fusion algorithms. Here, we propose a novel sensor bias correction method to correct the input data of the spatiotemporal fusion model through a machine learning technique learning the bias between different sensors. Taking the normalized difference vegetation index (NDVI) images with low-spatial resolution (MODIS) and high-spatial resolution (Landsat) as the basic data, we generated the neighborhood gray matrices from the MODIS image and established the image bias pairs of MODIS and Landsat. The light gradient boosting machine (LGBM) regression model was used for the nonlinear fitting of the bias pairs to correct MODIS NDVI images. For three different landscape areas with various spatial heterogeneities, the fusion of the bias-corrected MODIS NDVI and Landsat NDVI was conducted by using the spatiotemporal adaptive reflection fusion model (STARFM) and the flexible spatiotemporal data fusion method (FSDAF), respectively. The results show that the sensor bias correction method can enhance the spatially detailed information in the input data, significantly improve the accuracy and robustness of the spatiotemporal fusion technology, and extend the applicability of the spatiotemporal fusion models. [ABSTRACT FROM AUTHOR]

Accurate extraction of crop cultivated area and spatial distribution is essential for food security. Crop classification methods based on machine learning and deep learning and remotely sensed time-series data are widely utilized to detect crop planting area. However, few studies assess the effectiveness of machine learning and deep learning algorithm integrated time-series satellite data for identifying multiple crop type classification over a large-scale region. Hence, this study aims to evaluate the effectiveness of machine learning and deep learning models in crop classification and provide a framework for large-scale multiple crop type classification based on time-series of satellite data. The time-series of the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and NaE (combined NDVI and EVI) were adopted as input features, and four widely used machine learning models, including Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), and their integration (Stacking) were compared to examine the performance of multiple crop types (i.e., wheat, corn, wheat-corn, early rice, and early rice-late rice) classification in the North China Plain. The performance of two types of deep learning classifiers: the One-dimensional Convolutional Neural Network (Conv1D) and Long Short-Term Memory Networks (LSTM), were also tested. The results showed that the NaE feature performed best among three input features, and the Stacking model produced the highest accuracy (77.12%) compared to other algorithms. [ABSTRACT FROM AUTHOR]

Rapid urbanization causes potential changes in the urban landscape, resulting in significant changes in land surface temperature and outdoor thermal comfort. This urban growth has a detrimental impact on the health and comfort of residents. The comfort level experienced in any given region depends on various parameters, including atmospheric temperature, relative humidity, land use and land cover (LULC). In this study, we aim to examine the spatial variation of outdoor thermal comfort in the city of Hyderabad. To achieve this, we utilize medium-resolution Landsat 8 imageries along with in situ meteorological data. The classification of LULC is carried out using the maximum likelihood method. A machine learning tool known as Support Vector Machine (SVM) is implemented, with seven environmental indices such as normalized difference vegetation index (NDVI), normalized difference water index (NDWI), new built-up index (NBI), LST, brightness, greenness, and wetness to predict outdoor thermal comfort (THI). The study reveals significant variations in THI across different land covers. Barren lands exhibit the highest mean THI values (27.3), followed by built-up areas (26.9), vegetation (24.1), and water bodies (20.7). These findings indicate that barren and built-up areas are associated with higher levels of discomfort, while vegetated regions and water bodies provide more neutral to moderate comfort conditions. These results also highlight distinct spatial variations in THI across different regions of the city, demonstrating the influence of the urban landscape on outdoor thermal comfort. This research is vital for identifying specific areas within cities that require targeted mitigation strategies to enhance outdoor comfort. [ABSTRACT FROM AUTHOR]

Soil moisture (SM) is an indicator of the moisture status of the land surface, which is useful for monitoring extreme weather events. Representative global SM datasets include the National Aeronautics and Space Administration (NASA) Soil Moisture Active Passive (SMAP), the Global Land Data Assimilation System (GLDAS), and the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis 5 (ERA5), but due to their low spatial resolutions, none of these datasets well describe SM changes in local areas, and they tend to have a low accuracy. Machine learning (ML)-based SM predictions have demonstrated high accuracy, but obtaining semi-real-time SM information remains challenging, and the dependence of the validation accuracy on the data sampling method used, such as random or yearly sampling, has led to uncertainties. In this study, we aimed to develop an ML-based model for real-time SM estimation that can capture local-scale variabilities in SM and have reliable accuracy, regardless of the sampling method. This study was conducted in South Korea, and satellite image data, numerical weather prediction (NWP) data, and topographic data provided within one day were used as the input data. For SM modeling, 13 input variables affecting the surface SM status were selected: 10- and 20-day cumulative standardized precipitation indexes (SPI10 and SPI20), a normalized difference vegetation index (NDVI), downward shortwave radiation (DSR), air temperature (Tair), land surface temperature (LST), soil temperature (Tsoil), relative humidity (RH), latent heat flux (LE), slope, elevation, topographic ruggedness index (TRI), and aspect. Then, SM models based on random forest (RF) and automated machine learning (AutoML) were constructed, trained, and validated using random sampling and leave-one-year-out (LOYO) cross-validation. The RF- and AutoML-based SM models had significantly high accuracy rates based on comparisons with in situ SM (mean absolute error (MAE) = 2.212–4.132%; mean bias error (MBE) = −0.110–0.136%; root mean square error (RMSE) = 3.186–5.384%; correlation coefficient (CC) = 0.732–0.913), while the AutoML-based SM model tended to have a higher accuracy than the RF-based SM model, regardless of the data sampling method used. In addition, when compared to in situ SM data, the SM models demonstrated the highest accuracy, outperforming both GLDAS and ERA5 SM data and well representing changes in the dryness/wetness of the land surface according to meteorological events (heatwave, drought, and rainfall). The SM models proposed in this study can, thus, offer semi-real-time SM data, aiding in the monitoring of moisture changes in the land surface, as well as short-term meteorological disasters, like flash droughts or floods. [ABSTRACT FROM AUTHOR]

The quantification of forage availability in tropical grasses is generally done in a destructive and time-consuming manner, involving cutting, weighing, and waiting for drying. To expedite this process, non-destructive methods can be used, such as unmanned aerial vehicles (UAVs) equipped with high-definition cameras, mobile device images, and the use of the normalized difference vegetation index (NDVI). However, these methods have been underutilized in tropical pastures. A literature review was conducted to present the current state of remote tools' use in predicting forage availability and quality in tropical pastures. Few publications address the use of non-destructive methods to estimate forage availability in major tropical grasses (Megathyrsus maximus; Urochloa spp.). Additionally, these studies do not consider the fertility requirements of each cultivar and the effect of management on the phenotypic plasticity of tillers. To obtain accurate estimates of forage availability and properly manage pastures, it is necessary to integrate remote methods with in situ collection of soil parameters. This way, it will be possible to train machine learning models to obtain precise and reliable estimates of forage availability for domestic ruminant production. [ABSTRACT FROM AUTHOR]

Many applications in agriculture as well as other related fields including natural resources, environment, health, and sustainability, depend on recent and reliable cropland maps. Cropland extent and intensity plays a critical input variable for the study of crop production and food security around the world. However, generating such variables manually is difficult, expensive, and time consuming. In this work, we discuss a cost effective, fast, and simple machine-learning-based approach to provide reliable cropland mapping model using satellite imagery. The study includes four test regions, namely Iran, Mozambique, Sri-Lanka, and Sudan, where Sentinel-2 satellite imagery were obtained with assigned NDVI scores. The solution presented in this paper discusses a complete pipeline including data collection, time series reconstruction, and cropland extent and crop intensity mapping using machine learning models. The approach proposed managed to achieve high accuracy results ranging between 0.92 and 0.98 across the four test regions at hand. [ABSTRACT FROM AUTHOR]

• A new method of constructing hybrid model is used to retrieve Cyclone Global Navigation Satellite System (CYGNSS) soil moisture. • Bayesian model averaging (BMA) can reduce the model uncertainty. • Novel vegetation data of Sun-induced fluorescence (SIF) achieve better performance than Normalized Difference Vegetation Index (NDVI). Using Global Navigation Satellite System-Reflectometry (GNSS-R) for soil moisture (SM) retrieval has recently gained importance due to its high temporal-spatial resolution. However, the current methods, i.e., constructing a single machine learning (ML)-based model, have large model uncertainty resulting from ML networks and input schemes. Moreover, traditional Normalized Difference Vegetation Index (NDVI) cannot capture the rapid vegetation changes well. In this paper, a new SM retrieval method of constructing a hybrid model based on Bayesian model averaging (BMA) is employed to reduce the model uncertainty. Meanwhile, novel Sun-induced fluorescence (SIF) data is used as ancillary data to represent the rapid change of vegetation. We validate the proposed method at point and regional scales using in-situ data and the Global Land Data Assimilation System (GLDAS) product. The results demonstrate that our method has high accuracy and low uncertainty in SM retrieval. At the point scale, as accuracy indices, the average R (μ R) of BMA increases from 0.90 to 0.93 and the average root-mean-square-error (μ RMSE ) decreases from 0.034 c m 3 / c m 3 to 0.029 c m 3 / c m 3 ; as indices of uncertainty, the standard deviations of R and RMSE (σ R and σ RMSE ) decrease by 32 % and 9 % compared to the single ML-based model. For the regional scale, the μ R increases from 0.79 to 0.81, the μ RMSE decreases from 0.024 c m 3 / c m 3 to 0.023 c m 3 / c m 3 , and the σ R decreases by 19 %. Moreover, we take the point-scale experiment as an example for comparison to compare the performance of SIF with that of NDVI. The μ R of BMA trained by SIF is 0.03 higher than that trained by NDVI and the μ RMSE decreases by 0.002 c m 3 / c m 3 ; σ R and σ RMSE decrease by 25 % and 6 %. Based on these results, the proposed method can reduce the uncertainty and the advantage of SIF has potential for improving the SM retrieval. [ABSTRACT FROM AUTHOR]

Accurate estimation of leaf area index (LAI) is essential for precision agriculture, yet traditional ground-based measurements are destructive, time-consuming, and limited in scale. This study aimed to address the need for rapid, non-destructive LAI monitoring over large areas by evaluating unmanned aerial vehicle (UAV) data and machine learning (ML) models. A field experiment with four irrigation treatments was conducted to obtain wide range of LAI values over two years. Multispectral and thermal UAV images were acquired throughout the growing season along with destructive LAI measurements. Five ML algorithms, including K-Nearest Neighbors (K-NN), Extra Trees Regressor (ETR), eXtreme Gradient Boosting (XGBoost), Random Forest (RF), and Support Vector Regression (SVR) were tested. Also, feature selection procedures were implemented to obtain useful information among the features used for the ML model. Results showed the K-NN model achieved the highest accuracy (R2 = 0.97, RMSE = 0.46, MAE = 0.197), followed by ETR. The analysis of feature selection revealed that the combination of Normalized Difference Vegetation Index (NDVI) and canopy height (N D V I × H c) product had the highest importance, followed by Soil-Adjusted Vegetation Index (SAVI) and Green Normalized Difference Vegetation Index (GNDVI). Also, utilizing vegetation indices calculated from multiple spectral bands proved to be more effective than using individual bands alone. Overall, the study demonstrates that UAV data and ML techniques can estimate sorghum LAI precisely to support precision agriculture applications. Moreover, using low-cost UAVs equipped with multispectral sensors presents a cost-effective and reliable method for LAI estimation using ML models. • UAV data and machine learning were used for accurate LAI estimation in sorghum. • NDVI_Hc (NDVIxHc) is key for LAI estimation. • Vegetation indices were found to have greater importance than single band. • Five different machine learning models were tested for LAI estimation. [ABSTRACT FROM AUTHOR]

Aboveground biomass (AGB) data are important for profitable and sustainable pasture management. In this study, we hypothesized that vegetation indexes (VIs) obtained through analysis of moderate spatial resolution satellite data (Landsat‐8 and Sentinel‐2) and meteorological data can accurately predict the AGB of Brachiaria (syn. Urochloa) pastures in Brazil. We used AGB field data obtained from pastures between 2015 and 2019 in four distinct regions of Brazil to evaluate (i) the relationship between three different VIs—normalized difference vegetation index (NDVI), enhanced vegetation index 2 (EVI2) and optimized soil adjusted vegetation index (OSAVI)—and meteorological data with pasture aboveground fresh biomass (AFB), aboveground dry biomass (ADB) and dry‐matter content (DMC); and (ii) the performance of simple linear regression (SLR), multiple linear regression (MLR) and random forest (RF) algorithms for the prediction of pasture AGB based on VIs obtained through satellite imagery combined with meteorological data. The results highlight a strong correlation (r) between VIs and AGB, particularly NDVI (r = 0.52 to 0.84). The MLR and RF algorithms demonstrated high potential to predict AFB (R2 = 0.76 to 0.85) and DMC (R2 = 0.78 to 0.85). We conclude that both MLR and RF algorithms improved the biomass prediction accuracy using satellite imagery combined with meteorological data to determine AFB and DMC, and can be used for Brachiaria (syn. Urochloa) AGB prediction. Additional research on tropical grasses is needed to evaluate different VIs to improve the accuracy of ADB prediction, thereby supporting pasture management in Brazil. [ABSTRACT FROM AUTHOR]

The prediction of crop yield plays a crucial role in national economic development, encompassing grain storage, processing, and grain price trends. Employing multiple sensors to acquire remote sensing data and utilizing machine learning algorithms can enable accurate, fast, and nondestructive yield prediction for maize crops. However, current research heavily relies on single-type remote sensing data and traditional machine learning methods, resulting in the limited robustness of yield prediction models. To address these limitations, this study introduces a field-scale maize yield prediction model named the convolutional neural network–attention–long short-term memory network (CNN-attention-LSTM) model, which utilizes multimodal remote sensing data collected by multispectral and light detection and ranging (LIDAR) sensors mounted on unmanned aerial vehicles (UAVs). The model incorporates meteorological data throughout the crop reproductive stages and employs the normalized difference vegetation index (NDVI), normalized difference red edge (NDRE), soil-adjusted vegetation index (SAVI), and enhanced vegetation index (EVI) for the initial part of the vegetative stage (initial part of the V period), the later part of the vegetative stage (later part of the V period), the reproductive stage (R period), and the maturity stage (M period), along with LIDAR data for Point75–100 in the later part of the V period, Point80–100 in the R period, and Point50–100 in the M period, complemented by corresponding meteorological data as inputs. The resulting yield estimation demonstrates exceptional performance, with an R2 value of 0.78 and an rRMSE of 8.27%. These results surpass previous research and validate the effectiveness of multimodal data in enhancing yield prediction models. Furthermore, to assess the superiority of the proposed model, four machine learning algorithms—multiple linear regression (MLR), random forest regression (RF), support vector machine (SVM), and backpropagation (BP)—are compared to the CNN-attention-LSTM model through experimental analysis. The outcomes indicate that all alternative models exhibit inferior prediction accuracy compared to the CNN-attention-LSTM model. Across the test dataset within the study area, the R2 values for various nitrogen fertilizer levels consistently exceed 0.75, illustrating the robustness of the proposed model. This study introduces a novel approach for assessing maize crop yield and provides valuable insights for estimating the yield of other crops. [ABSTRACT FROM AUTHOR]

As more machine learning and deep learning models are applied in studying the quantitative relationship between the climate and terrestrial vegetation growth, the uncertainty of these advanced models requires clarification. Partial dependence plots (PDPs) are one of the most widely used methods to estimate the marginal effect of independent variables on the predicted outcome of a machine learning model, and it is regarded as the main basis for conclusions in relevant research. As more controversies regarding the reliability of the results of the PDPs emerge, the uncertainty of the PDPs remains unclear. In this paper, we experiment with real, remote sensing data to systematically analyze the uncertainty of partial dependence relationships between four climate variables (temperature, rainfall, radiation, and windspeed) and vegetation growth, with one conventional linear model and six machine learning models. We tested the uncertainty of the PDP curves across different machine learning models from three aspects: variation, whole linear trends, and the trait of change points. Results show that the PDP of the dominant climate factor (mean air temperature) and vegetation growth parameter (indicated by the normalized difference vegetation index, NDVI) has the smallest relative variation and the whole linear trend of the PDP was comparatively stable across the different models. The mean relative variation of change points across the partial dependence curves of the non-dominant climate factors (i.e., radiation, windspeed, and rainfall) and vegetation growth ranged from 8.96% to 23.8%, respectively, which was much higher than those of the dominant climate factor and vegetation growth. Lastly, the model used for creating the PDP, rather than the relative importance of these climate factors, determines the fluctuation of the PDP output of these climate variables and vegetation growth. These findings have significant implications for using remote sensing data and machine learning models to investigate the quantitative relationships between the climate and terrestrial vegetation. [ABSTRACT FROM AUTHOR]

Evapotranspiration (ET) plays an important role in transferring water and converting energy in the land–atmosphere system. Accurately estimating ET is crucial for understanding global climate change, ecological environmental problems, the water cycle, and hydrological processes. Machine learning (ML) algorithms have been considered as a promising method for estimating ET in recent years. However, due to the limitations associated with the spatial–temporal resolution of the flux tower data commonly used as the target set in ML algorithms, the ability of ML to discover the inherent laws within the data is reduced. In this study, a hybrid framework was established to simulate ET in data-deficient areas. ET simulation results of a coupled model comprising the Budyko function and complementary principle (BC2021) were used as the target set of the random forest model, instead of using the flux station observation data. By combining meteorological and hydrological data, the monthly ET of the Inner Mongolia section of the Yellow River Basin (IMSYRB) was simulated from 1982 to 2020, and good results were obtained (R2 = 0.94, MAE = 3.82 mm/mon, RMSE = 5.07 mm/mon). Furthermore, the temporal and spatial variations in ET and the influencing factors were analysed. In the past 40 years, annual ET in the IMSYRB ranged between 241.38 mm and 326.37 mm, showing a fluctuating growth trend (slope = 0.80 mm/yr), and the summer ET accounted for the highest proportion in the year. Spatially, ET in the IMSYRB showed a regular distribution of high ET in the eastern region and low ET in the western area. The high ET value areas gradually expanded from east to west over time, and the area increased continuously, with the largest increase observed in the 1980s. Temperature, precipitation, and normalized difference vegetation index (NDVI) were found to be the most important factors affecting ET in the region and play a positive role in promoting ET changes. These results provide an excellent example of long-term and large-scale accurate ET simulations in an area with sparse flux stations. [ABSTRACT FROM AUTHOR]