Accurate estimation of biomass and Leaf Area Index (LAI) requires appropriate models and predictor variables. These biophysical parameters are indicative of crop productivity, and thus, are of interest in applications such as crop yield forecasting and precision farming. This study evaluated the potential of leveraging vegetation indices derived from multi-temporal RapidEye data using a machine learning approach to estimate crop biomass and LAI. Both near-infrared and red-edge based indices were considered in this study. In-situ measurements of these two parameters for three main cash crops, including canola, corn, and soybeans, were collected during a field campaign and used for model calibration and validation. Crops models were developed using the artificial neural network (ANN) and support vectors regression (SVR). Results showed that, for each crop, the SVR modeled LAI and biomass more accurately than ANN. For biomass, the SVR’s Root Mean Square Errors (RMSEs) were reported as 25.22 g/m² for canola, 88.13 g/m² for corn, 5.91 g/m² for soybean, and 56.14 g/m² for all crops pooled. Similarly, for the LAI, SVR provided the best model with RMSE = 0.59 m²/m² for canola, RMSE = 0.27 m²/m² for corn, RMSE = 0.21 m²/m² for soybean, and RMSE = 0.51 m²/m² for all crops together.

The water cloud model (WCM) can be inverted to estimate leaf area index (LAI) using the intensity of backscatter from synthetic aperture radar (SAR) sensors. Published studies have demonstrated that the WCM can accurately estimate LAI if the model is effectively calibrated. However, calibration of this model requires access to field measures of LAI as well as soil moisture. In contrast, machine learning (ML) algorithms can be trained to estimate LAI from satellite data, even if field moisture measures are not available. In this study, a support vector machine (SVM) was trained to estimate the LAI for corn, soybeans, rice, and wheat crops. These results were compared to LAI estimates from the WCM. To complete this comparison, in situ and satellite data were collected from seven Joint Experiment for Crop Assessment and Monitoring (JECAM) sites located in Argentina, Canada, Germany, India, Poland, Ukraine and the United States of America (U.S.A.). The models used C-Band backscatter intensity for two polarizations (like-polarization (VV) and cross-polarization (VH)) acquired by the RADARSAT-2 and Sentinel-1 SAR satellites. Both the WCM and SVM models performed well in estimating the LAI of corn. For the SVM, the correlation (R) between estimated LAI for corn and LAI measured in situ was reported as 0.93, with a root mean square error (RMSE) of 0.64 ${\mathrm{m}}^2{\mathrm{m}}^{-2}$ and mean absolute error (MAE) of 0.51 ${\mathrm{m}}^2{\mathrm{m}}^{-2}$. The WCM produced an R-value of 0.89, with only slightly higher errors (RMSE of 0.75 ${\mathrm{m}}^2{\mathrm{m}}^{-2}$ and MAE of 0.61 ${\mathrm{m}}^2{\mathrm{m}}^{-2}$) when estimating corn LAI. For rice, only the SVM model was tested, given the lack of soil moisture measures for this crop. In this case, both high correlations and low errors were observed in estimating the LAI of rice using SVM (R of 0.96, RMSE of 0.41 ${\mathrm{m}}^2{\mathrm{m}}^{-2}$ and MAE of 0.30 ${\mathrm{m}}^2{\mathrm{m}}^{-2}$). However, the results demonstrated that when the calibration points were limited (in this case for soybeans), the WCM outperformed the SVM model. This study demonstrates the importance of testing different modeling approaches over diverse agro-ecosystems to increase confidence in model performance.

Biomass has a direct relationship with agricultural production and may help to predict crop yield. Earth observation technology can contribute significantly to monitoring given the availability of temporally frequent and high-resolution radar or optical satellite data. Polarimetric Synthetic Aperture Radar (PolSAR) has several advantages for operational monitoring given that at these longer wavelengths atmospheric and illumination conditions do not affect acquisitions and considering the sensitivity of microwaves to the structural properties of targets. Therefore, SARs are a promising source of data for crop mapping and monitoring. With increasing access to SARs the development of robust methods to monitor crop productivity is timely. In this paper, we examine the use of machine learning and artificial intelligence approaches to analyze a time series of Polarimetric parameters for crop biomass estimation. In total, 14 polarimetric parameters from a time series of Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) airborne L-band data were used for biomass estimation for an intensively cropped site in western Canada. Then, Multiple linear regression (MR) and artificial neural network (ANN) models were developed and evaluated to estimate the biomass for canola, corn, and soybeans. According to the experimental results, the ANN provided more accurate biomass estimates compared to MR. Canola biomass, in general, showed less sensibility to almost all the polarimetric parameters. Nevertheless, Freeman-Double combined with vertical-vertical backscattering (VV) delivered the correlation coefficient (r) of 0.72, and the root mean square error (RMSE) of 56.55 g m⁻²of canola biomass. For corn, the highest correlation was observed between a pairing of horizontal- horizontal backscattering (HH) with Entropy (H) for biomass estimation yielding an r of 0.92 and RMSE of 196.71 g m⁻². Horizontal-vertical backscattering (HV) and Yamaguchi-Surface (O_Y) delivered the highest sensitivity for soybeans (r of 0.82 and RMSE of 13.48 g m⁻²). If all crops are pooled, H combined with O_Y provided the most accurate estimates of biomass (r of 0.89 and RMSE of 135.31 g m⁻²). These results demonstrated that models which make use of polarimetric parameters that characterize the multiple sources of scattering typical of vegetation canopies can be used to estimate crop biomass accurately. Such results bode well for agricultural monitoring considering the increasing number of satellite SAR sensors with various frequencies, imaging modes and revisit times. As such, the time series analysis and methods proposed in this study could be used to monitor crop development and productivity using SAR space technologies.

On 10 August 2020, a series of intense and fast-moving windstorms known as a derecho caused widespread damage across Iowa’s (the top US corn-producing state) agricultural regions. This severe weather event bent and flattened crops over approximately one-third of the state. Immediate evaluation of the disaster’s impact on agricultural lands, including maps of crop damage, was critical to enabling a rapid response by government agencies, insurance companies, and the agricultural supply chain. Given the very large area impacted by the disaster, satellite imagery stands out as the most efficient means of estimating the disaster impact. In this study, we used time-series of Sentinel-1 data to detect the impacted fields. We developed an in-season crop type map using Harmonized Landsat and Sentinel-2 data to assess the impact on important commodity crops. We intersected a SAR-based damage map with an in-season crop type map to create damaged area maps for corn and soybean fields. In total, we identified 2.59 million acres as damaged by the derecho, consisting of 1.99 million acres of corn and 0.6 million acres of soybean fields. Also, we categorized the impacted fields to three classes of mild impacts, medium impacts and high impacts. In total, 1.087 million acres of corn and 0.206 million acres of soybean were categorized as high impacted fields.

Polarimetric decomposition extracts scattering features that are indicative of the physical characteristics of the target. In this study, three polarimetric decomposition methods were tested for soil moisture estimation over agricultural fields using machine learning algorithms. Features extracted from model-based Freeman–Durden, Eigenvalue and Eigenvector based H/A/α, and Van Zyl decompositions were used as inputs in random forest and neural network regression algorithms. These algorithms were applied to retrieve soil moisture over soybean, wheat, and corn fields. A time series of polarimetric Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) data acquired during the Soil Moisture Active Passive Experiment 2012 (SMAPVEX12) field campaign was used for the training and validation of the algorithms. Three feature selection methods were tested to determine the best input features for the machine learning algorithms. The most accurate soil moisture estimates were derived from the random forest regression algorithm for soybeans, with a correlation of determination (R²) of 0.86, root mean square error (RMSE) of 0.041 m³ m⁻³ and mean absolute error (MAE) of 0.030 m³ m⁻³. Feature selection also impacted results. Some features like anisotropy, Horizontal transmit and Horizontal receive (HH), and surface roughness parameters (correlation length and RMS-H) had a direct effect on all algorithm performance enhancement as these parameters have a direct impact on the backscattered signal.

Crop biophysical parameters, such as Leaf Area Index (LAI) and biomass, are essential for estimating crop productivity, yield modeling, and agronomic management. This study used several features extracted from multi-temporal Sentinel-1 Synthetic Aperture Radar (SAR) and spectral vegetation indices extracted from Sentinel-2 optical data to estimate crop LAI and wet and dry biomass. Various machine learning algorithms, including Random Forest Regression (RFR), Support Vector Regression (SVR), and Artificial Neural Network (ANN), were trained and assessed for three major crops (wheat, soybeans and canola). ANN provided the best accuracy for all wheat parameters and soybean LAI and canola wet biomass and LAI. RFR led to higher accuracy for soybean dry and wet biomass. However, SVR could accurately estimate only canola dry biomass. All data were then pooled to investigate if a single algorithm could estimate biophysical parameters for all crops. The RFR model accurately estimated wet and dry biomass and LAI across all crop types in this scenario. This generic model is fast and accurate and can be easily applied for crop mapping and monitoring over large geographies using cloud computing platforms, such as Google Earth Engine.