189 research outputs found
Remote Sensing of Evapotranspiration for Operational Drought Monitoring Using Principles of Water and Energy Balance
No full textEvapotranspiration (ET) is an important component of the hydrologic budget because it reflects the exchange of mass and energy between the soil–water–vegetation system and the atmosphere. Prevailing weather conditions influence potential or reference ET through variables such as radiation, temperature, wind, and relativity humidity. In addition to these weather variables, actual ET (ETa) is also affected by land cover type and condition, as well as soil moisture. The dependence of ETa on land cover and soil moisture, and its direct relationship with carbon dioxide assimilation in plants, makes it an important variable for monitoring drought, crop yield, and biomass—a critical capability for decision makers interested in food security, grain markets, water allocation, and carbon sequestration (Bastiaanssen et al., 2005).
Because ET can be difficult to measure accurately, especially at large spatial scales, several different hydrologic modeling techniques have been developed to estimate ETa using satellite remote sensing. In general, the ET modeling techniques can be grouped into two broad classes that include models based on surface energy balance (e.g., Bastiaanssen et al., 1998; Su et al., 2005; Allen et al., 2007; Anderson et al., 2007; Senay et al., 2007) and water balance (e.g., Allen et al., 1998, Senay, 2008) principles. While water balance models focus on tracking the pathways and magnitude of rainfall in the soil–vegetation system, most remote sensing energy balance models use land surface temperature (LST) as a primary constraint in partitioning radiant energy available at the surface between heat and water fluxes
Agricultural drought projection in Ethiopia from 1981 to 2050: Using coordinated regional climate downscaling experiment climate data for Africa
No full textModeling Landscape Evapotranspiration by Integrating Land Surface Phenology and a Water Balance Algorithm
No full textThe main objective of this study is to present an improved modeling technique called Vegetation ET (VegET) that integrates commonly used water balance algorithms with remotely sensed Land Surface Phenology (LSP) parameter to conduct operational vegetation water balance modeling of rainfed systems at the LSP’s spatial scale using readily available global data sets. Evaluation of the VegET model was conducted using Flux Tower data and two-year simulation for the conterminous US. The VegET model is capable of estimating actual evapotranspiration (ETa) of rainfed crops and other vegetation types at the spatial resolution of the LSP on a daily basis, replacing the need to estimate crop- and region-specific crop coefficients
Fusion of In-Situ and Remote Sensing Data with Machine Learning Toward Global Water and Food Security Monitoring
No full textWater and food insecurity are two of the most formidable challenges facing the world today: an estimated two billion people (26%) are without safe drinking water and four billion people (50%) contend with water scarcity for at least part of the year while two and a half billion people (30%) regularly experience moderate to severe food insecurity. The United Nations Sustainable Development Goals (SDGs) are designed to promote and monitor progress across several dimensions of economic, social, and environmental sustainability, including water and food security. Unfortunately, global progress towards these targets ahead of the 2030 deadline is severely off track. Moreover, global estimates tend to obscure inequalities between and within countries and low- and middle-income populations in sub-Saharan Africa are faring the worst due to the impact of climate change and greater socioeconomic vulnerability. An urgent call for action and policy solutions is warranted to identify what needs to be done, for and with whom. However, current monitoring schemes for many SDG targets, including those related to water and food security, are not up to the challenge and exhibit problems of data latency and coarse resolution at odds with the type of data and decision tools that would enable targeted solutions. Many researchers, policymakers, and practitioners are advocating for the localization of the SDGs and programmatic monitoring to realize real progress. Given data limitations in most developing contexts, satellite remote sensing and the use of machine learning tools may be one of the only cost-effective means of providing high-resolution, high frequency data that is uniform, collected in near-real time, and available across all geographies. The data fusion of existing in-situ data and satellite-derived observations, enabled by machine learning, complements ongoing monitoring efforts and advances the localization of monitoring.
This dissertation presents how remote sensing, in-situ data, and machine learning can be assimilated to produce novel insights in service of localized monitoring and evidence-based targeting to address water and food insecurity in the arid and semi-arid lands (ASALs) of Kenya and Rwanda.
In the Kenyan ASALs, population-based groundwater use and demand at 5 km resolution was predicted during dry season months for years 2017-2021 with in-season accuracy of 70-76% and forecasted accuracy of 68-75%. This tool assimilated both in-situ water pumping and remote sensing data and presented the first operational spatially explicit sub-seasonal to seasonal (S2S) maps of groundwater use in the literature. Knowledge of historical and forecasted groundwater use is designed to improve decision making and resource allocation for drought early warning and contribute to monitoring of indicators for SDG 6 related to water access and use by pastoralists.In Rwanda, wall-to-wall time series of maize area and yield at 10 m resolution and for 10 maize-growing seasons from 2019-2023 were developed to study the agricultural productivity of smallholders. Informed by in-situ crop cutting data, maize was classified with 83% accuracy and mean district-level yields had a root mean square error (RMSE) of 370 kg/ha (27%) against national estimates. This dataset enabled the evaluation of village-level response to a rural transportation project and the monitoring of Rwanda's recent growth in agricultural productivity as part of SDG 2 to achieve food security. While national progress is off track to double maize yields by 2030, this growth is unequal across villages and producers, and, depending on policy priorities, there are several scenarios for meeting national goals, bringing up the lowest yielding farmers, or both within the next six years.
High-resolution data, enabled by machine learning and in-situ data, identified previously unmeasured spatial and temporal variation in indicators of water and food insecurity in these two contexts. It is likely that inequalities and inconsistent progress in attainment of the SDGs would be found for many other regions and SDG indicators at high-resolution. Highly differentiated and targeted policies and programs can accelerate progress and reduce inequalities efficiently. Thus, the localization of monitoring may be one necessary condition to achieve the SDGs fully in the short time remaining before 2030.</p
Fusion of In-Situ and Remote Sensing Data with Machine Learning Toward Global Water and Food Security Monitoring
No full textWater and food insecurity are two of the most formidable challenges facing the world today: an estimated two billion people (26%) are without safe drinking water and four billion people (50%) contend with water scarcity for at least part of the year while two and a half billion people (30%) regularly experience moderate to severe food insecurity. The United Nations Sustainable Development Goals (SDGs) are designed to promote and monitor progress across several dimensions of economic, social, and environmental sustainability, including water and food security. Unfortunately, global progress towards these targets ahead of the 2030 deadline is severely off track. Moreover, global estimates tend to obscure inequalities between and within countries and low- and middle-income populations in sub-Saharan Africa are faring the worst due to the impact of climate change and greater socioeconomic vulnerability. An urgent call for action and policy solutions is warranted to identify what needs to be done, for and with whom. However, current monitoring schemes for many SDG targets, including those related to water and food security, are not up to the challenge and exhibit problems of data latency and coarse resolution at odds with the type of data and decision tools that would enable targeted solutions. Many researchers, policymakers, and practitioners are advocating for the localization of the SDGs and programmatic monitoring to realize real progress. Given data limitations in most developing contexts, satellite remote sensing and the use of machine learning tools may be one of the only cost-effective means of providing high-resolution, high frequency data that is uniform, collected in near-real time, and available across all geographies. The data fusion of existing in-situ data and satellite-derived observations, enabled by machine learning, complements ongoing monitoring efforts and advances the localization of monitoring.
This dissertation presents how remote sensing, in-situ data, and machine learning can be assimilated to produce novel insights in service of localized monitoring and evidence-based targeting to address water and food insecurity in the arid and semi-arid lands (ASALs) of Kenya and Rwanda.
In the Kenyan ASALs, population-based groundwater use and demand at 5 km resolution was predicted during dry season months for years 2017-2021 with in-season accuracy of 70-76% and forecasted accuracy of 68-75%. This tool assimilated both in-situ water pumping and remote sensing data and presented the first operational spatially explicit sub-seasonal to seasonal (S2S) maps of groundwater use in the literature. Knowledge of historical and forecasted groundwater use is designed to improve decision making and resource allocation for drought early warning and contribute to monitoring of indicators for SDG 6 related to water access and use by pastoralists.In Rwanda, wall-to-wall time series of maize area and yield at 10 m resolution and for 10 maize-growing seasons from 2019-2023 were developed to study the agricultural productivity of smallholders. Informed by in-situ crop cutting data, maize was classified with 83% accuracy and mean district-level yields had a root mean square error (RMSE) of 370 kg/ha (27%) against national estimates. This dataset enabled the evaluation of village-level response to a rural transportation project and the monitoring of Rwanda's recent growth in agricultural productivity as part of SDG 2 to achieve food security. While national progress is off track to double maize yields by 2030, this growth is unequal across villages and producers, and, depending on policy priorities, there are several scenarios for meeting national goals, bringing up the lowest yielding farmers, or both within the next six years.
High-resolution data, enabled by machine learning and in-situ data, identified previously unmeasured spatial and temporal variation in indicators of water and food insecurity in these two contexts. It is likely that inequalities and inconsistent progress in attainment of the SDGs would be found for many other regions and SDG indicators at high-resolution. Highly differentiated and targeted policies and programs can accelerate progress and reduce inequalities efficiently. Thus, the localization of monitoring may be one necessary condition to achieve the SDGs fully in the short time remaining before 2030.</p
Using High Spatial Resolution Spectral Data To Study Spatial and Temporal Variability in Corn and Soybean Management Systems /
No full textComparison of Four Different Energy Balance Models for Estimating Evapotranspiration in the Midwestern United States
No full textThe development of different energy balance models has allowed users to choose a model based on its suitability in a region. We compared four commonly used models—Mapping EvapoTranspiration at high Resolution with Internalized Calibration (METRIC) model, Surface Energy Balance Algorithm for Land (SEBAL) model, Surface Energy Balance System (SEBS) model, and the Operational Simplified Surface Energy Balance (SSEBop) model—using Landsat images to estimate evapotranspiration (ET) in the Midwestern United States. Our models validation using three AmeriFlux cropland sites at Mead, Nebraska, showed that all four models captured the spatial and temporal variation of ET reasonably well with an R2 of more than 0.81. Both the METRIC and SSEBop models showed a low root mean square error (<0.93 mm·day−1) and a high Nash–Sutcliffe coefficient of efficiency (>0.80), whereas the SEBAL and SEBS models resulted in relatively higher bias for estimating daily ET. The empirical equation of daily average net radiation used in the SEBAL and SEBS models for upscaling instantaneous ET to daily ET resulted in underestimation of daily ET, particularly when the daily average net radiation was more than 100 W·m−2. Estimated daily ET for both cropland and grassland had some degree of linearity with METRIC, SEBAL, and SEBS, but linearity was stronger for evaporative fraction. Thus, these ET models have strengths and limitations for applications in water resource management
Land Cover Change Effects on Stormflow Characteristics across Broad Hydroclimate Representative Urban Watersheds in the United States
No full textUrban development alters stormflow characteristics and is associated with increasing flood risks. The long-term evaluation of stormflow characteristics that exacerbate floods, such as peak stormflow and time-to-peak stormflow at varying levels of urbanization across different hydroclimates, is limited. This study investigated the long-term (1980s to 2010s) effects of increasing urbanization on key stormflow characteristics using observed 15 min streamflow data across six broad hydroclimate representative urban watersheds in the conterminous United States. The results indicate upward trends in peak stormflow and downward trends in time-to-peak stormflow at four out of six watersheds. The watershed in the Great Plains region had the largest annual increasing (decreasing) percent change in peak stormflow (time-to-peak stormflow). With the current change rates, peak stormflow in the Great Plains region watershed is expected to increase by 55.4% and have a 2.71 h faster time-to-peak stormflow in the next decade
Evaluation of streamflow predictions from LSTM models in water- and energy-limited regions in the United States
No full textThe application of Long Short-Term Memory (LSTM) models for streamflow predictions has been an area of rapid development, supported by advancements in computing technology, increasing availability of spatiotemporal data, and availability of historical data that allows for training data-driven LSTM models. Several studies have focused on improving the performance of LSTM models; however, few studies have assessed the applicability of these LSTM models across different hydroclimate regions. This study investigated the single-basin trained local (one model for each basin), multi-basin trained regional (one model for one region), and grand (one model for several regions) models for predicting daily streamflow in water-limited Great Basin (18 basins) and energy-limited New England (27 basins) regions in the United States using the CAMELS (Catchment Attributes and Meteorology for Large-sample Studies) data set. The results show a general pattern of higher accuracy in daily streamflow predictions from the regional model when compared to local or grand models for most basins in the New England region. For the Great Basin region, local models provided smaller errors for most basins and substantially lower for those basins with relatively larger errors from the regional and grand models. The evaluation of one-layer and three-layer LSTM network architectures trained with 1-day lag information indicates that the addition of model complexity by increasing the number of layers may not necessarily increase the model skill for improving streamflow predictions. Findings from our study highlight the strengths and limitations of LSTM models across contrasting hydroclimate regions in the United States, which could be useful for local and regional scale decisions using standalone or potential integration of data-driven LSTM models with physics-based hydrological models
Characterizing Crop Water Use Dynamics in the Central Valley of California Using Landsat-Derived Evapotranspiration
No full textUnderstanding how different crops use water over time is essential for planning and managing water allocation, water rights, and agricultural production. The main objective of this paper is to characterize the spatiotemporal dynamics of crop water use in the Central Valley of California using Landsat-based annual actual evapotranspiration (ETa) from 2008 to 2018 derived from the Operational Simplified Surface Energy Balance (SSEBop) model. Crop water use for 10 crops is characterized at multiple scales. The Mann–Kendall trend analysis revealed a significant increase in area cultivated with almonds and their water use, with an annual rate of change of 16,327 ha in area and 13,488 ha-m in water use. Conversely, alfalfa showed a significant decline with 12,429 ha in area and 13,901 ha-m in water use per year during the same period. A pixel-based Mann–Kendall trend analysis showed the changing crop type and water use at the level of individual fields for all of Kern County in the Central Valley. This study demonstrates the useful application of historical Landsat ET to produce relevant water management information. Similar studies can be conducted at regional and global scales to understand and quantify the relationships between land cover change and its impact on water use
- …
