پايش و پيش‌بيني تغييرات كاربري اراضي حوضۀ آبخيز شهرستان فارسان با استفاده از مدل LCM

Monitoring and Predicting of Land Use Changes in Farsan Watershed Using LCM

ﭘﻮﺷـﺶ اراﺿـﯽ ﺣﻮﺿـﻪﻫﺎي آﺑﺨﯿﺰ از ﻣﺒﺎﺣﺚ ﺗﻮﺳـﻌﻪاي ﻣﺘﺄﺛﺮ و ﺑﻪ ﺷـﺪت در ﺣﺎل ﺗﻐﯿﯿﺮ ﻫﺴـﺘﻨﺪ ﮐﻪ اﯾﻦ ﺗﻐﯿﯿﺮات ﺑﺮ ﻣﺘﻐﯿﺮﻫﺎي دﯾﮕﺮي ﻫﻤﭽﻮن ﻫﯿﺪروﻟﻮژي ﺣﻮ ﺿﮥ آﺑﺨﯿﺰ ﺗﺄﺛﯿﺮ ﮔﺬارﻧﺪ. ﻫﺪف اﯾﻦ ﺗﺤﻘﯿﻖ، ﭘﺎﯾﺶ ﺗﻐﯿﯿﺮات ﮐﺎرﺑﺮي ارا ﺿﯽ در ﮔﺬﺷﺘﻪ و ﺑﺮرﺳﯽ اﻣﮑﺎن ﭘﯿﺶﺑﯿﻨﯽ آن در آﯾﻨﺪه ﺑﺎ اﺳﺘﻔﺎده از ﻣﺪلﺳﺎزي ﺗﻐﯿﯿﺮ زﻣﯿﻦ )LCM1( در ﺣﻮﺿﮥ آﺑﺨﯿﺰ ﺷﻬﺮﺳﺘﺎن ﻓﺎرﺳﺎن اﺳﺘﺎن ﭼﻬﺎرﻣﺤﺎلوﺑﺨﺘﯿﺎري اﺳﺖ. ﺑﺪﯾﻦ ﻣﻨﻈﻮر، ﺗﺼﺎوﯾﺮ ﺳﻨﺠﻨﺪهﻫﺎي ﻟﻨﺪﺳﺖ-5 TM ﺑﺮاي ﺳﺎلﻫﺎي 1365 و 1388 و ﺗﺼﺎوﯾﺮ ﺳﻨﺠﻨﺪة ﻟﻨﺪﺳﺖ-8 OLI ﺑﺮاي ﺳﺎل 1396 ﻣﻮرد ﺗﺠﺰﯾﻪ و ﺗﺤﻠﯿﻞ ﻗﺮار ﮔﺮﻓﺖ. ﻃﺒﻘﺎت ﮐﺎرﺑﺮي ارا ﺿﯽ ﺷﺎﻣﻞ ﻣﻨﺎﻃﻖ ﻣ ﺴﮑﻮﻧﯽ، ارا ﺿﯽ ﮐ ﺸﺎورزي، ﮐ ﺸﺖ دﯾﻢ، ﻣﺮاﺗﻊ، ﺳﻨﮕﻼخﻫﺎ، ﭘﻬﻨﻪﻫﺎي آﺑﯽ، ﺧﺎك ﻟﺨﺖ و ﺑﺮف در ﺗﺼﺎوﯾﺮ ﻫﺮ ﺳﻪ دورة زﻣﺎﻧﯽ ﻃﺒﻘﻪﺑﻨﺪي ﺷﺪ. ﭘﯿﺶﺑﯿﻨﯽ وﺿﻌﯿﺖ ﮐﺎرﺑﺮي اراﺿﯽ ﺑﺮاي ﺳﺎل 1396 ﺑﻪ ﮐﻤﮏ ﻣﺪل LCM ﺑﺮ ﭘﺎﯾﮥ ﺷﺒﮑﮥ ﻋ ﺼﺒﯽ ﻣ ﺼﻨﻮﻋﯽ و ﺗﺤﻠﯿﻞ زﻧﺠﯿﺮة ﻣﺎرﮐﻮف ﻧﯿﺰ اﻧﺠﺎم ﮔﺮﻓﺖ. ﭘﺲ از ارزﯾﺎﺑﯽ ﺻﺤﺖ ﻣﺪل ﺑﺎ ا ﺳﺘﻔﺎده از آﻣﺎرة ﮐﺎﭘﺎ، ﻧﻘ ﺸﮥ ﭘﻮ ﺷﺶ ارا ﺿﯽ ﺳﺎل 1406 ﺑﺎ ا ﺳﺘﻔﺎده از دورة وا ﺳﻨﺠﯽ 1365- 1396 ﭘﯿﺶﺑﯿﻨﯽ ﺷﺪ. ﻧﺘﺎﯾﺞ ﻧ ﺸﺎن داد ﮐﻪ ارا ﺿﯽ ﻣﺮﺗﻌﯽ ﻃﯽ ﺳﺎلﻫﺎي 1365 ﺗﺎ 1396 ﺑﻪ ﻣﯿﺰان 4379 ﻫﮑﺘﺎر ﮐﺎﻫﺶ ﭘﯿﺪا ﮐﺮدهاﻧﺪ وﻟﯽ ارا ﺿﯽ ﮐ ﺸﺎورزي ﺑﻪ ﻣﯿﺰان 1922 ﻫﮑﺘﺎر اﻓﺰوده ﺷﺪهاﻧﺪ. در اﯾﻦ ﻣﻄﺎﻟﻌﻪ، LCM ﺗﻮاﻧ ﺴﺖ 85% از ﺗﻐﯿﯿﺮات را ﺑﻪ درﺳﺘﯽ ﭘﯿﺶﺑﯿﻨﯽ ﻧﻤﺎﯾﺪ. در ﺳﺎل 1406 ﺑﻪ ﻣﯿﺰان 149ﻫﮑﺘﺎر اﻓﺰاﯾﺶ وﺳﻌﺖ اراﺿﯽ ﺷﻬﺮي و 100 ﻫﮑﺘﺎر ﮐﺎﻫﺶ ارا ﺿﯽ ﻣﺮﺗﻌﯽ ﻣﻨﻄﻘﻪ ﭘﯿﺶﺑﯿﻨﯽ ﺷﺪ. ﻟﺬا ﺿﻤﻦ ﺗﺄﮐﯿﺪ ﺑﺮ ﺣﻔﻆ ارا ﺿﯽ ﻣﺮﺗﻌﯽ، ﻻزم ا ﺳﺖ از اﯾﻦ ﺗﮑﻨﯿﮏ ﺑﺮاي ﭘﯿﺶﺑﯿﻨﯽ ﺗﻐﯿﯿﺮات، ﻋﻠﻞ آن و ﻫﻤﭽﻨﯿﻦ ﺗﺒﻌﺎت ﺗﻐﯿﯿﺮات ﮐﺎرﺑﺮي اراﺿﯽ در ﺳﻄﻮح وﺳﯿﻊﺗﺮي ﺑﺮرﺳﯽ و ﺑﺮﻧﺎﻣﻪرﯾﺰي ﮐﺮد.

The land use of the watersheds is one of the most affected and highly vulnerable due to developmental process which effect on the other variables such as the hydrological function. The purpose of this research is to monitor land use changes in the past and to investigate predictability of its future using Land Change Modeler (LCM) in the watershed of Farsan County of Chaharmahal-va-Bakhtiari province. For this purpose, the Landsat-5 TM images of 1986 and 2009 as well as the Landsat-8 OLI images of 2017 were analyzed. Land covers including residential areas, agricultural land, dryland farming, rangelands, rocks, water bodies, bare-land and snow were classified for the three periods. The prediction of land cover of 2017 was done using the LCM model based on Artificial Neural Network and Markov chain analysis after assessing model’s accuracy based on Kappa index. The land cover of 2027 was also predicted using a change probability table extracted from occurred changes from 1986-2017. The results show that the rangeland decreased by 4379-ha in the years 1986 to 2017, but the agricultural land increased by 1922-ha. This study proved that the LCM could accurately forecast future changes (85% overall accuracy). An increase of 149-ha of residential area and 100-ha decrease of rangelands in the study area was predicted for 2027. Therefore, while emphasizing the conservation of rangelands, it is necessary to study and use this technique to predict changes, its causes, as well as the consequences of land use changes at the broader scales.