Introducing dryland furrow soil erosion; a new form of water erosion and investigating the dependency on soil properties in the semi-arid regions, NW Iran.

[Display omitted] • Soil loss from the furrows under rainfall is a result of interaction between interrill and rill erosion processes. • Dryland Furrow Soil Erosion (DFSE) is introduced as a new form of water erosion. • DFSE is affected by soil particle size distribution, structure, organic matter and permeability. • Soil particle size distribution is the most factor determining the DFSE in rainfed lands. Soil erosion by water is the most serious form of land degradation in dry farming lands of semi-arid regions where tillage are mostly done up and down slope which accelerate water erosion processes in the soil. Cultivated furrows are similar to rills and transport soil particles detached by interrill and rill processes during heavy rainfalls. Soil loss from the furrows is an interaction of interrill and rill erosion processes. This can be introduced as dryland furrow soil erosion (DFSE). There is limited information on the factors influencing the DFSE in semi-arid regions. A laboratory experiment was set up for ten dominant soil textures (clay loam, sandy loam, sandy clay loam, loam, loamy sand, sandy clay, silty clay loam, silty clay, silt loam and clay) in erosion flume (60 cm wide and 400 cm long) under simulated rainfalls with 50 mm h−1 for 60 min. Thirty DFSE and soil properties data were randomly divided to two groups consist of twenty data for modeling DFSE and ten data for validation of the model. The DFSE values varied from 0.014 g m−2 s−1 to 0.092 g m−2 s−1 with an average of 0.054 g m−2 s−1. The DFSE was negatively correlated with sand (r = -0.95), organic matter (r = -0.59), MWD of water-stable aggregates (r = -0.49) and infiltration rate (r = -0.46) and positively correlated with silt (r = 0.71), clay (r = 0.72), dispersible clay (r = 0.77). Multiple linear regression analysis using twenty data showed that the DFSE is related to silt and clay content (R2 = 0.92, p < 0.001). Evaluation of the model showed that the developed model has lower error with a mean absolute error of 9.11 % and a model efficiency of 0.92. Therefore, the model can be reliably used for predicting the DFSE in the similar areas in soil and rainfalls in semi-arid regions. [ABSTRACT FROM AUTHOR]