The emergence of remote sensing technologies coupled with local monitoring workstations enables us the unprecedented ability to monitor the environment in large scale. Information mining from multi-channel geo-spatiotemporal data however poses great challenges to many computational sustainability applications. Most existing approaches adopt various dimensionality reduction techniques without fully taking advantage of the spatiotemporal nature of the data. In addition, the lack of labeled training data raises another challenge for modeling such data. In this work, we propose a novel semi-supervised attention-based deep representation model that learns context-aware spatiotemporal representations for prediction tasks. A combination of convolutional neural networks with a hybrid attention mechanism is adopted to extract spatial and temporal variations in the geo-spatiotemporal data. Recognizing the importance of capturing more complete temporal dependencies, we propose the hybrid attention mechanism which integrates a learnable global query into the classic self-attention mechanism. To overcome the data scarcity issue, sampled spatial and temporal context that naturally reside in the largely-available unlabeled geo-spatiotemporal data are exploited to aid meaningful representation learning. We conduct experiments on a large-scale real-world crop yield prediction task. The results show that our methods significantly outperforms existing state-of-the-art yield prediction methods, especially under the stress of training data scarcity.

Surface roughness parameterization plays an important role in passive microwave soil moisture (SM) retrieval. This article proposes a new formulation for estimating surface roughness. The proposed model incorporates the field-scale (micro) roughness as well as topographic (macro) roughness. The performance of the model is evaluated by inverting the traditional tau-omega model for retrieving SM. The study focuses on the passive active L-band system (PALS) radiometer data collected as a part of two Soil Moisture Active Passive Validation Experiment (SMAPVEX), i.e., SMAPVEX12 (humid Manitoba, Canada) and SMAPVEX15 (semiarid Arizona, USA) with highly different microroughness and macroroughness. The measured surface roughness is observed to increase exponentially with clay fraction (CF). This behavior is minimized with increase in leaf area index (LAI). In the absence of vegetation, the contribution of topography toward surface roughness increases. A higher surface roughness value is estimated for SMAPVEX12, which positively correlate with LAI and CF and negatively correlate with wetness conditions. On the other hand, due to the high topographic variability in SMAPVEX15 region, the contribution of topography (surface curvature) toward total surface roughness is significant. Also, consistently dry SM resulted in high microroughness for SMAPVEX15. Nevertheless, a total surface roughness estimated for SMAPVEX15 region is less than for SMAPVEX12. The surface roughness formulation presented in this study can be extrapolated to any spatial resolution.

The Soil Moisture Active Passive (SMAP) mission currently provides three surface soil moisture products based solely on instrument measurements. The three soil moisture products are: 1) the radiometer-only 36 km gridded; 2) a radiometer-only enhanced product gridded at 9 km; and 3) a high-resolution (3 km) SMAP-Sentinel active–passive product. It is important to validate these released SMAP soil moisture products over various land covers and hydroclimatic domains before they are routinely used in scientific research and applications. This paper evaluates SMAP-based soil moisture products for typical Indian conditions of extreme seasonal variability that leads to changes from very wet to dry soil, especially for the paddy dominated region. The assessment metrics indicate that the enhanced passive-only soil moisture product meets the SMAP accuracy requirement of 0.04 m3/m3during the nongrowing season (NGS) with unbiased root-mean-square error (ubRMSE) values ranging between 0.025 and 0.036 m3/m3. However, this product underperformed during the paddy growing season (GS) with ubRMSE values ranging between 0.063 and 0.097 m3/m3. In addition, the SMAP-Sentinel active–passive soil moisture product shows satisfactory performance during the NGS (ubRMSE, 0.017–0.051 m3/m3), but during the GS, ubRMSE ranged between 0.089 and 0.104 m3/m3. Use of the vegetation water content climatology and low clay fraction in SMAP baseline algorithm (auxiliary database) that mismatched with the actual values may be the possible source of errors and biases in the SMAP soil moisture products. The reported study provides guidelines for the application of enhanced SMAP soil moisture products in India, especially for the tropical region, and provides information that can be used to improve the retrieval algorithm.

The SMOS, and SMAP retrievals are not only influenced by soil moisture, vegetation but also highly sensitive to surface roughness. The current retrieval algorithms use roughness models and parameterization developed at field scale to estimate soil moisture at satellite footprint. These models ignored large scale roughness features due to differences in elevation, concavity etc., observed within the satellite footprint. In this study, it is hypothesized that the satellite observed surface roughness is contributed by micro-roughness (small-scale roughness) and macro-roughness (large-scale roughness), incorporating the scattering features observed at field and satellite scale. The traditional algorithm is used to retrieve soil moisture for SMAPVEX12 (Soil Moisture Active Passive Validation Experiment, 2012), and SMAPVEX15 the field experiments.

Satellite-based remote sensing of soil moisture is generally conducted with active (radar) and passive (radiometer) microwave measurements. During active microwave remote sensing the backscattering from the target i.e, the soil surface is adversely affected by the overlaying vegetation, consequently, sending degraded signal back to the radar sensor. This phenomena greatly compromise with the quality of soil moisture measurements. The proposed research presents an algorithm that averts usage of theoretical and empirical backscattering models. The algorithm uses Soil-Vegetation-Atmosphere-Transfer model for soil moisture estimation that is used to quantify the backscattering components of radar signals. The algorithm has simple and valid assumptions that convert the total radar backscattering equations for a particular temporal scale into a set of simple linear systems. The algorithm reasonably estimates the stochastic surface roughness and vegetation backscattering components.