3.1 Potential of UAV imagery to study soil moisture and stock of water in wetland
The selection of the modeling approach depends, besides the modelling objectives, on both scientific and technical aspects as well as on the resources available, which include the scale of the simulation domain, topography, hydrological and topographical data available and the complexity of the hydraulic regime (Wester et al., 2018). For these reasons, to estimate water storage the studied wetland was subdivided into three compartments, in order to facilitate the hydrodynamic modeling. The compartments were defined through rigorous analysis of the images (RGB and DMS) generated by UAV, which showed distinct pixel-information about soil surface inside that vary according topography. Thus, we considered the topography and terrain morphology: the slope, from the border to the center of the topographic depression that characterizes the wetland site, and the density of the microtopography (mound fields), which is directly related to the level of hydromorphism in the wetland (Rosolen et al., 2019).
Between the border and the center of the depression, compartment 1 corresponds to its boundary with the well-drained plateau; compartment 2 is the intermediate segment between compartments 1 (boundary) and 3 (center); and compartment 3 corresponds to the central part of the topographic depression (Figure 2 C).
The data acquired by the UAV images allowed us to estimate the amount of water that each compartment can support and, therefore, the total water volume that can be stored on the surface of the studied wetland. Based on the data of the altitude, geographical location and the slope in relation to adjacent points of each pixel of the images we calculated the volume by delineating a 3D polygon in each compartment. Because the mounds (murundus ) do not fill up with water, their volumes were subtracted from the total volume of the compartments.
The result of semi-automatic extraction through slope and visual interpretation of the orthomosaic obtained 1751 mounds features in the research area. The biggest mound have 3643.40 m² and the smallest mapped shape was 1.01 m². Regarding the perimeter recorded in the extraction, the smallest recorded form was 4.60 m and the largest with 476.63 m. Figure 6 shows the histograms with murundus distribution by area and perimeter in the research area. It is observed that the murundus of the research area are mostly distributed with perimeters ranging from 24.48 m to 58.49 m and the area are distributed mostly between 67.41 m² and 167.37 m².
When each compartment is flooded, it displays a distinct water residence time and dynamic associated with its topographic position. Water flooding over the land surface reaches compartment 1, during the rainfalls. In compartment 3 the water level remains high for a prolonged period due the lateral water influx from compartment 2 and 1. Clearly, there are variation in soil moisture pattern linked to topography, and that variable has a major impact on the hydrology in landscape (Moore et al., 1991). The largest amount of microtopography with little trees and shrubs on their top occurs in Compartment 1 in agreement with lower soil moisture. Wetland area also exhibit grassy cover in Compartment 3 in agreement with the elevated soil moisture. These differences of soil moisture observed in field are in accordance to features observed in the aerial imaging.
Rainfall seasonality determines the wetland’s water supply. In the field, we observed that the wetland area is a natural seasonal water reservoir, filling up in rainy months and drying up during dry months (Figure 7 A, B). We performed the photogrammetry at dry months in order to estimate the water stock during the rainy period.
Table 2 shows each compartment’s total storage capacity, in cubic meters and liters (Figure 2 D), its area in square meters, and the minimum and maximum altitudes that were used in the calculations. The altitude values ​​were calculated for each pixel of the images of each compartment.
In compartment 3, we conducted the electrical tomography tests to provide an understanding of subsurface hydrodynamics and identify the architecture of soil and rock, which acts as an infiltration zone.