Abstract. This study investigates the ability of machine learning models to retrieve the surface soil moisture of a grassland area from multispectral remote sensing carried out using an unoccupied aircraft system (UAS). In addition to multispectral images, we use terrain attributes derived from a digital elevation model and hydrological variables of precipitation and potential evapotranspiration as covariates to predict surface soil moisture. We tested four different machine learning algorithms and interrogated the models to rank the importance of different variables and to understand their relationship with surface soil moisture. All the machine learning algorithms we tested were able to predict soil moisture with good accuracy. The boosted regression tree algorithm was marginally the best, with a mean absolute error of 3.8 % volumetric moisture content. Variable importance analysis revealed that the four most important variables were precipitation, reflectance in the red wavelengths, potential evapotranspiration, and topographic position indices (TPI). Our results demonstrate that the dynamics of soil water status across heterogeneous terrain may be adequately described and predicted by UAS remote sensing and machine learning. Our modeling approach and the variable importance and relationships we have assessed in this study should be useful for management and environmental modeling tasks where spatially explicit soil moisture information is important.

No abstract

One important challenge for Sandhill Crane conservation is the collection of regular, accurate counts of birds roosting in flooded areas at night. Estimates of roost site numbers are an important way to track population size over time and detect any changes in site selection that may be a function of management or changes in site conditions. Traditional methods for estimating roosting crane numbers are morning and evening counts of cranes as they arrive at or depart their roosting sites. The accuracy of these counts however is often hampered by poor visibility in foggy and/or low light conditions, variations between observers, difficulties in accurately estimating the number of birds when many are flying to or away from the site, and difficulties in pinpointing specific locations where birds are roosting. Use of small unmanned aerial systems (sUAS) equipped with infrared cameras provide a promising alternative for developing more accurate estimates of roosting population numbers efficiently and more frequently, as well as a method of mapping specific roosting locations relative to habitat features. sUAS could also assist in surveying new areas for roosting cranes, as even though these birds are often traditional, they are known to colonize new areas if habitat conditions are suitable. This paper presents a case study on how to get approval from the Federal Aviation Administration (FAA) for night flight missions, how to decide the mission parameters and timing, as well as initial post-processing workflows. From 11 night missions performed, we also will share in this paper our findings and lessons learned.

We developed machine learning models to retrieve surface soil moisture (0-4 cm) from high resolution multispectral imagery, terrain attributes, and local climate covariates. Using a small unmanned aircraft system (UAS) equipped with a multispectral sensor we captured high resolution imagery in part to create a high-resolution digital elevation model (DEM) as well as quantify relative vegetation photosynthetic status. We tested four different machine learning algorithms. The boosted regression tree algorithm provided the best accuracy model with mean absolute error of 3.8 % volumetric water content. The most important variables for the prediction of soil moisture were precipitation, reflectance in the red wavelengths, potential evapotranspiration, and topographic position indices (TPI). Our results demonstrate that the dynamics of soil water status across heterogeneous terrain may be adequately described and predicted by UAS remote sensing data and machine learning. Our modeling approach and the variable importance and relationships we have assessed in this study should be useful for management and environmental modeling tasks where spatially explicit soil moisture information is important.