Multi-source machine learning and spaceborne remote sensing data accurately predict three-dimensional soil moisture in an in-service uranium disposal cell.

One reason arid and semi-arid environments have been used to store waste is due to low groundwater recharge, presumably limiting the potential for meteoric water to mobilize and transport contaminants into groundwater. The U.S. Department of Energy Office of Legacy Management (LM) is evaluating selected uranium mill tailings disposal cell covers to be managed as evapotranspiration (ET) covers, where vegetation is used to naturally remove water from the cover profile via transpiration, further reducing deep percolation. An important parameter in monitoring the performance of ET covers is soil moisture (SM). If SM is too high, water may drain into tailings material, potentially transporting contaminants into groundwater; if SM is too low, radon flux may increase through the cover. However, monitoring SM via traditional instrumentation is invasive, expensive, and may fail to account for spatial heterogeneity, especially over vegetated disposal cells. Here we investigated the potential for non-invasive SM monitoring using radar remote sensing and other geospatial data to see if this approach could provide a practical, accurate, and spatially comprehensive tool to monitor SM. We used theoretical simulations to analyze the sensitivity of multi-frequency radar backscatter to SM at different depths of a field-scale (3 ha) drainage lysimeter embedded within an in-service LM disposal cell. We then evaluated a shallow and deep form of machine learning (ML) using Google Earth Engine to integrate multi-source observations and estimate the SM profile across six soil layers from depths of 0-2 m. The ML models were trained using in situ SM measurements from 2019 and validated using data from 2014 to 2018 and 2020-2021. Model predictors included backscatter observations from satellite synthetic aperture radar, vegetation, temperature products from optical infrared sensors, and accumulated, gridded rainfall data. The radar simulations confirmed that the lower frequencies (L- and P-band) and smaller incidence angles show better sensitivity to deeper soil layers and an overall larger SM dynamic range relative to the higher frequencies (C- and X-band). The ML models produced accurate SM estimates throughout the soil profile (r values from 0.75 to 0.94; RMSE = 0.003-0.017 cm3/cm3; bias = 0.00 cm3/cm3), with the simpler shallow-learning approach outperforming a selected deep-learning model. The ML models we developed provide an accurate, cost-effective tool for monitoring SM within ET covers that could be applied to other vegetated disposal cell covers, potentially including those with rock-armored covers.

Application and Comparison of the MODIS-Derived Enhanced Vegetation Index to VIIRS, Landsat 5 TM and Landsat 8 OLI Platforms: A Case Study in the Arid Colorado River Delta, Mexico.

The Enhanced Vegetation Index (EVI) is a key Earth science parameter used to assess vegetation, originally developed and calibrated for the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard the Terra and Aqua satellites. With the impending decommissioning of the MODIS sensors by the year 2020/2022, alternative platforms will need to be used to estimate EVI. We compared Landsat 5 (2000⁻2011), 8 (2013⁻2016) and the Visible Infrared Imaging Radiometer Suite (VIIRS; 2013⁻2016) to MODIS EVI (2000⁻2016) over a 420,083-ha area of the arid lower Colorado River Delta in Mexico. Over large areas with mixed land cover or agricultural fields, we found high correspondence between Landsat and MODIS EVI (R² = 0.93 for the entire area studied and 0.97 for agricultural fields), but the relationship was weak over bare soil (R² = 0.27) and riparian vegetation (R² = 0.48). The correlation between MODIS and Landsat EVI was higher over large, homogeneous areas and was generally lower in narrow riparian areas. VIIRS and MODIS EVI were highly similar (R² = 0.99 for the entire area studied) and did not show the same decrease in performance in smaller, narrower regions as Landsat. Landsat and VIIRS provide EVI estimates of similar quality and characteristics to MODIS, but scale, seasonality and land cover type(s) should be considered before implementing Landsat EVI in a particular area.

NDVI, scale invariance and the modifiable areal unit problem: An assessment of vegetation in the Adelaide Parklands.

This research addresses the question as to whether or not the Normalised Difference Vegetation Index (NDVI) is scale invariant (i.e. constant over spatial aggregation) for pure pixels of urban vegetation. It has been long recognized that there are issues related to the modifiable areal unit problem (MAUP) pertaining to indices such as NDVI and images at varying spatial resolutions. These issues are relevant to using NDVI values in spatial analyses. We compare two different methods of calculation of a mean NDVI: 1) using pixel values of NDVI within feature/object boundaries and 2) first calculating the mean red and mean near-infrared across all feature pixels and then calculating NDVI. We explore the nature and magnitude of these differences for images taken from two sensors, a 1.24m resolution WorldView-3 and a 0.1m resolution digital aerial image. We apply these methods over an urban park located in the Adelaide Parklands of South Australia. We demonstrate that the MAUP is not an issue for calculation of NDVI within a sensor for pure urban vegetation pixels. This may prove useful for future rule-based monitoring of the ecosystem functioning of green infrastructure.