Assessment of sedimentation process in flood water spreading system using IRS (P5) and supervised classification algorithms (case study: Dahandar plain, Minab city, south of Iran)

Abstract

Many parts of Iran are dry plains that seem suitable places for building flood spreading systems and diverting flash floods to infiltration areas in order to recharge ground water resources. After passing a period of time, sediment deposition decreases the artificial recharge capacity of flood spreading system infiltration basins. Panchromatic images of IRS(P5) satellite are appropriate tools for studying sediment distribution, texture and thickness on the surface of these systems. For monitoring sediment distribution, the panchromatic images used that were taken in 2007,2010 and 2016. For evaluating sediment texture and thickness, 40 samples were collected from surface soil of different places all over Dahandar flood spreading system, in Minab city. By using Hydrometer method and soil texture triangle, different percentages of sand, silt, clay, soil texture types and sediment thickness were determined. Then by using simple linear regression the relationship between different percentages of sand, silt, clay and sediment thickness with the digital numbers of 2016 Panchromatic image were evaluated. The values of R2 for sand, silt, clay and sediment thickness were 0.428, 0.343, 0.227 and 0.429 respectively. Different algorithms of supervised classification (maximum likelihood, minimum distance, parallelepiped) were used for creating maps with different classes of sand, silt, clay, soil texture types and sediment thickness. For sand particles, maximum likelihood algorithm with over all accuracy of 64%, for silt particles minimum distance-raw algorithm with over all accuracy of 51%, for clay particles minimum distance-normalized with over all accuracy of 46% showed the greatest values. For sediment thickness, different algorithms had the same function. The mentioned methods could reasonably explain the distribution of sand, silt and clay particles, soil texture and sediment thickness all over the flood spreading system.