It is widely known that soil erosion is an issue of concern in soil and water quality, affecting agriculture and natural resources. Thus, scientific efforts must take into consideration the high-resolution elevation dataset in order to implement a precision conservation approach effectively. New advances such as LiDAR products have provided a basic source of information to enable researchers to identify small erosional landscape features. To fill this gap, this study developed a methodology based on data mining of hydrologic and topographic attributes associated with concentrated flow path identification to distinguish classic gully side walls and bed areas. At 0.91 Km2 region of the Keigley Branch-South Skunk River watershed, an area with gullies, we computed profile curvature, mean slope deviation, stream power index, and aspect gridded in 1-m pixel from Iowa LiDAR project. CLARA (CLustering LARge Applications) algorithm. An unsupervised clustering approach was employed on 913,495 points splitting the dataset in six groups, the number in agreement with within-group sum of squared error (WSS) statistical technique. In addition, a new threshold criteria termed gully concentrated flow (GCF) based upon data distribution of flow accumulation and mean overall slope were introduced to produce polylines that identified the main hydrographic flow paths, corresponding to the gully beds. Cluster #6 was classified as gully side walls. After distinguishing gullies and cliffs areas among points belonging to cluster 6, all six gullies were satisfactorily identified. The proposed methodology improves on existent techniques because identifies distinct parts of gullies which include side walls and bed zone.

Another important concept is assessing gully slope stability in order to generate useful information for precision conservation planning. Although limit-equilibrium concept has been used widely in rock mechanics its application in precision conservation structures is relatively new. This study evaluated two multi-temporal surveys in a Western Iowa gullied area under the approach of soil stability regarding precision conservation practice The study computed factor of safety (FS) at the gully area, including headcut and gully side walls using digital elevation models originated from surveys conducted in 1999 and 2014.

Outcomes of this assessment have revealed significantly less instability of the actual slopes compared to 1999 survey slopes. The internal friction angle (θ) had the largest effect on slope stability factor (S.D.1999 = 0.18, S.D.2014 = 0.24), according the sensitivity analysis, compared to variations of soil cohesion, failure plane angle and slab thickness. In addition, critically instable slopes within gully, based on units of the slope standard deviation, as a threshold, have produced an area of 61 m2 and 396 m2 considering the threshold of one and two slope standard deviation, respectively. The majority of these critical areas were located near the headcut and in the border of side walls. Based on current literature, association of processed material (geotextile) and crop cover with high root density might be an alternative to improve slope instability, but empirical tests are necessary to validate this approach. Nevertheless, the slope instability must include other factors that capture the dynamics of failure.