Tracking ephemeral gullies formation and development in agricultural conditions using the mathematical model AnnAGNPS

On cultivated land, ephemeral gullies contribute significantly to soil erosion. Despite their importance, existing models for predicting the location of ephemeral gully initiation and development are very sparse and have limitations. The USDA-ARS National Sedimentation Laboratory and the University of Nottingham developed GIS-based topographic analyses to map potential ephemeral gully locations. In the 1980s they proposed an indicator called Compound Topographic Index (CTI), which was proposed as a predictor of gully location, and which is defined, for a certain pixel located in a watershed, as the product of the watershed area at that point, the local slope and the local curvature. It can be seen that this product is a proxy for the power of the stream at that point. From a digital elevation model it is possible to calculate for each pixel the value of its CTI. Researchers at the aforementioned centers found that pixels exceeding a CTI value, called critical CTI, very often corresponded to the location of areas eroded by ephemeral gullies.

This work aims to test the suitability of the CTI-based method to locate the gullies observed in typical conditions of an agricultural plot as a starting point to evaluate the performance of the QAnnAGNPS model, which develops a whole technology based on this methodology. Although such methodology is old, evaluations of it in real agricultural situations are extraordinarily scarce. In addition, it is still necessary to verify in field conditions that the modeling approach based on headcut occurence and migration, localized by means of CTI, is correct.  Thus, an experiment has been initiated in November 2023 in which, first of all, an agricultural plot has been selected in an area of highly erodible silty loam soils located in Pitillas (Navarra). The plot has been tilled with conventional tillage to replicate the initial conditions of an average agricultural plot, which has been kept free of vegetation by using herbicide. After each precipitation event, drone flights have been carried out to obtain digital elevation models (DEM) with a resolution of less than one centimeter and orthomosaics. The DEMs and orthomosaics generated in each flight make it possible to locate the origin of the gullies formed and to determine their dimensions and their temporal evolution, in this case until November 2024, when the plot was tilled again to restart the observations.

 

These observed data were compared with the data simulated by QAnnAGNPS. For each gully, we first obtained the critical CTI by selecting the one that best explains the origin of the gully. On the other hand, for each gully and time, we have the CTI threshold value that best follows the gully's path.  It has been confirmed that CTI gives good results for locating ephemeral gullies in real agricultural conditions and is a good way to predict the path of a gully. Our observations confirm that ephemeral gullies were always formed from generation and migration upstream of headcut, so the theoretical basis of models such as AnnAGNPS is valid.