Multitemporal Analysis of Gully Erosion in Olive Groves by Means of Digital Elevation Models Obtained with Aerial Photogrammetric and LiDAR Data

Abstract

Gully erosion is one of the main processes of soil degradation, representing 50%–90% of total erosion at basin scales. Thus, its precise characterization has received growing attention in recent years. Geomatics techniques, mainly photogrammetry and LiDAR, can support the quantitative analysis of gully development. This paper deals with the application of these techniques using aerial photographs and airborne LiDAR data available from public database servers to identify and quantify gully erosion through a long period (1980–2016) in an area of 7.5 km2 in olive groves. Several historical flights (1980, 1996, 2001, 2005, 2009, 2011, 2013 and 2016) were aligned in a common coordinate reference system with the LiDAR point cloud, and then, digital surface models (DSMs) and orthophotographs were obtained. Next, the analysis of the DSM of differences (DoDs) allowed the identification of gullies, the calculation of the affected areas as well as the estimation of height differences and volumes between models. These analyses result in an average depletion of 0.50 m and volume loss of 85000 m3 in the gully area, with some periods (2009–2011 and 2011–2013) showing rates of 10,000–20,000 m3/year (20–40 t/ha*year). The manual edition of DSMs in order to obtain digital elevation models (DTMs) in a detailed sector has facilitated an analysis of the influence of this operation on the erosion calculations, finding that it is not significant except in gully areas with a very steep shape.

Analysis of the Evolution of Gully Erosion in Olive Groves Using Photogrammetry Techniques. Relationships With Rainfall Regime

“…Both aerial photogrammetric and LiDAR data are available from several public Spatial Data Infrastructures (SDI) and download services (Fernández et al, 2020a). Thus, photographs of the national flights were downloaded from the photo-library of the National Geographic Institute of Spain (IGN) and photographs of the regional flights were downloaded from the photo-library of Andalusia (Institute of Statistics and Cartography of Andalusia, IECA).…”

“…These points were measured in constructions and building roof corners, unequivocally identifiable in the LiDAR data. They were fixed by adjusting a flat surface to the LiDAR points located on the roofs and computing the minimum bounding rectangle (Fernández et al, 2020a). The mean X and Y errors calculated were lower than 0.20 m being the RMS and SD of about 0.5 m. Meanwhile, the mean vertical (Z) error was of 0.24 m, the SD was of 0.40 m and the RMS was of 0.46 m.…”

“…From these DSMs, grid files were exported to a GIS software (QGIS) and then transformed to raster in TIFF format. After Fernández et al(2016Fernández et al( , 2020a, DSMs were used due to the inconveniences to generate DTMs. The DSM uncertainties were the Z errors calculated at CHK computed in the orientation process (Fernandez et al, 2020a).…”

“…Besides, images can be taken from different platforms: terrestrial (Castillo et al, 2012;Eltner et al, 2013;Kaiser et al, 2014;Koci et al, 2017); unmanned aerial systems (UAS) (Eltner et al, 2013;Koci et al, 2017; aerial platforms (Brasington et al, 2003;Martínez-Casasnovas et al 2004;Fernández el al., 2020a); or satellites (Wang et al, 2016). From images, different approaches have been applied such as conventional photogrammetric techniques (Brasington et al, 2003;Martínez-Casasnovas et al 2004;Fernández et al, 2020a) or new computer vision techniques (Kaiser et al, 2014;Wang et al, 2016;Koci et al, 2017).…”

“…The availability of data at different epochs leads to multitemporal analysis of gully areas, either in 2D (Hayas et al, 2017) or 3D approaches, usually by means of the calculation of DEMs of Differences (DoDs). These methods are often combined with point capture techniques such as LiDAR, GNSS or TS (Brasington et al, 2003;Castillo et al, 2012;Koci et al, 2017;Fernández et al, 2020a). UAS in their different modalities are well suited for very high resolution and accurate surveys in areas between 0.01-100 km 2 .…”