What Drives Gullies in Spain's Olive Landscapes? A regional analysis of gully activity

Data of research work "What Drives Gullies in Spain's Olive Landscapes? A Regional Analysis of Gully Activity" Figure 1. Photographs of the main landscape units (a) Mid mountains (b) Valley plains (c) Countryside foothills (d) Countryside hills Figure 2 (a) Guadalquivir basin in Andalusia (Spain) (b) study area, one within each of the principal landscapes units occupied by olive groves and the meteorological stations (c) area occupied by olive groves in Andalusia (SIOSE) Figure 3. Graphic illustration of gully head activity over the three study periods, showing four scenarios (NAG is new active gully head, SG is stable gully head, EAG is existing active gully head, X indicates no gully head presence, NON). Scenario a) In 2008-2010 there is no gully head, in 2010-2013 a new gully head appears (NAG) and the gully head remains in the same position in 2013-2019 (SG); Scenario b) A gully head is mapped in the same position in all photos from 2008 to 2019, so the gully remains stable (SG) over all three study periods, Scenario c) A gully head emerges (NAG) in 2008-2010 and advances in 2010-2013 and in 2013-2019. This gully head is considered existing active gully head in the last two periods (EAG) d) A gully head emerges (NAG) in 2008-2010 and the gully head remains stable in 2010-2013 and the gully head advances in 2013-2019. Figure 4. Observed gully heads in the years 2008, 2010, 2013 and 2019 for each landscape unit: a) countryside foothills, b) countryside hills, c) mid mountains and d) valley plains. Each study area occupies 25 km2. Figure 5. Gully Head Initiation (GHI) index for the three gully head activity types and points that are not gully heads (EAG – existing active gully heads, SG - stable gully heads, NAG – new active gully heads, and NON- non-gully heads) in the 4 landscape units and for the entire data set (without differentiation between landscapes). The line dividing the box represents the median value. Figure 6. Comparing GHI across three study periods (EAG: existing active gully heads, SG: stable gully heads, NAG: new active gully heads, and NON: non-gully heads) Figure 7. (a) ROC curves and corresponding AUC values for each pairwise combination of the three gully head activity types, considering all landscapes together. (b) AUC for the entire model together, considering all the points that are gully heads compared to those that are not. 0: Non gully heads, NON; 1: New active gully heads NAG; 2: Existing active gully heads EAG; 3: Stable gully heads, SG. Figure 8. Topographic threshold for gully head initiation applied to the three classes of gully head erosion activity and points that are not gully heads (SG: Stable gully heads, NAG: New active gully heads, EAG: Existing active gully heads, NON: Non-gully head)