كارآيي مدل‌هاي نفوذ آب به خاك در كاربري‌هاي مختلف زمين در حوزه آبخيز تهم‌چاي

The Efficiency of Water Infiltration Models in Different Land Uses of the Tahamchai Catchment

ﭘﯿﺶﺑﯿﻨﯽ دﻗﯿﻖ ﺷﺪت ﻧﻔﻮذ آب ﺑﻪ ﺧﺎك ﺑﺮاي ﺑﺮﻧﺎﻣﻪرﯾﺰي ﺣﻔﺎﻇﺖ آب و ﺧﺎك در ﺣﻮزهﻫﺎي آﺑﺨﯿﺰ اﻫﻤﯿﺖ زﯾﺎدي دارد. در اﯾﻦ ﻣﻄﺎﻟﻌﻪ ﮐﺎرآﯾﯽ ﻣﺪلﻫﺎي ﻧﻔﻮذ در ﮐﺎرﺑﺮيﻫﺎي ﮐﺸﺖ دﯾﻢ، ﮐﺸﺖ آﺑﯽ و ﻣﺮﺗﻊ )20 ﻣﮑﺎن در ﺳﻪ ﺗﮑﺮار( در ﺣﻮزه آﺑﺨﯿﺰ ﺗﻬﻢ ﭼﺎي واﻗﻊ در ﺷﻤﺎل ﻏﺮﺑﯽ اﺳﺘﺎن زﻧﺠﺎن، اﯾﺮان، ﺑﺮرﺳﯽ ﺷﺪ. ﮐﺎراﯾﯽ ﻣﺪلﻫﺎي ﻧﻔﻮذ ﺑﺮ اﺳﺎس دادهﻫﺎي اﻧﺪازهﮔﯿﺮي ﺷﺪه ﺑﺎ اﺳﺘﻔﺎده از آﻣﺎرهﻫﺎي ﺿﺮﯾﺐ ﺗﻌﯿﯿﻦ )R2(، رﯾﺸﻪ ﻣﯿﺎﻧﮕﯿﻦ ﻣﺮﺑﻌﺎت ﺧﻄﺎ )RMSE( و ﺿﺮﯾﺐ ﮐﺎراﯾﯽ ﻧﺶ-ﺳﺎﺗﮑﻠﯿﻒ )NSE( ارزﯾﺎﺑﯽ ﺷﺪ. ﺑﺮ اﺳﺎس ﻧﺘﺎﯾﺞ، ﺗﻐﯿﯿﺮات آﺷﮑﺎري از ﻧﻈﺮ ﺷﺪت ﻧﻔﻮذ آب در ﮐﺎرﺑﺮيﻫﺎي ﻣﺨﺘﻠﻒ زﻣﯿﻦ وﺟﻮد دارد و ﺑﯿﺶﺗﺮﯾﻦ ﻣﻘﺪار در ﮐﺎرﺑﺮي دﯾﻢ )16/6ﺳﺎﻧﺘﯽﻣﺘﺮ ﺑﺮ ﺳﺎﻋﺖ( و ﮐﻤﺘﺮﯾﻦ آن در ﮐﺎرﺑﺮي ﻣﺮﺗﻊ )5/4 ﺳﺎﻧﺘﯽﻣﺘﺮ ﺑﺮ ﺳﺎﻋﺖ( اﺳﺖ. از ﺑﯿﻦ ﻣﺪلﻫﺎي ﻣﻮرد ﺑﺮرﺳﯽ، ﻣﺪل ﮐﻮﺳﺘﯿﺎﮐﻮف ﺑﯿﺸﺘﺮﯾﻦ ﺿﺮﯾﺐ ﺗﻌﯿﯿﻦ )R2=0/99( و ﮐﻤﺘﺮﯾﻦ ﻣﻘﺪار ﺧﻄﺎ )1-RMSE=0/05 cm h( و ﺑﯿﺶﺗﺮﯾﻦ ﺿﺮﯾﺐ ﮐﺎراﯾﯽ )0/99=NSE( را در ﮐﺎرﺑﺮي ﻣﺮﺗﻊ ﻧﺸﺎن داد. در ﮐﺎرﺑﺮي ﮐﺸﺖ دﯾﻢ و ﮐﺸﺖ آﺑﯽ ﻣﺪل ﻓﻠﯿﭗ ﺑﺎ ﺿﺮﯾﺐ ﺗﻌﯿﯿﻦ 0/99 و ﺑﺎ ﻣﯿﺎﻧﮕﯿﻦ ﻣﺮﺑﻌﺎت ﺧﻄﺎﯾﯽ از 0/14 ﺗﺎ 0/22 و ﻫﻤﭽﻨﯿﻦ ﺿﺮﯾﺐ ﮐﺎراﯾﯽ ﻧﺶ-ﺳﺎﺗﮑﻠﯿﻒ 0/99 ﻧﺴﺒﺖ ﺑﻪ دﯾﮕﺮ ﻣﺪلﻫﺎ ﮐﺎراﯾﯽ ﺑﯿﺸﺘﺮي را داﺷﺖ. ﺑﺎ وﺟﻮد آن ﮐﻪ ﻣﺪل ﮔﺮﯾﻦ-اﻣﭙﺖ ﺑﺮازش ﻧﺴﺒﺘﺎً ﺧﻮﺑﯽ ﺑﻪ دﯾﮕﺮ ﻣﺪلﻫﺎ ﺑﻪ دادهﻫﺎي اﻧﺪازهﮔﯿﺮي ﺷﺪه داﺷﺖ )R2=0/94( اﻣﺎ ﺷﺎﺧﺺ ﺧﻄﺎي آن ﻧﺴﺒﺘﺎً زﯾﺎد )RMSE=2/32 cm h1 ( ﺑﻮد ﮐﻪ ﻧﺸﺎن از ﮐﻢﺗﺮﯾﻦ ﮐﺎرآﯾﯽ اﯾﻦ ﻣﺪل در ﺑﯿﻦ ﻣﺪلﻫﺎ اﺳﺖ. ﺧﻄﺎي اﯾﻦ ﻣﺪل در ﻣﻘﺎدﯾﺮ زﯾﺎدﺗﺮ ﺷﺪت ﻧﻔﻮذ ﮐﻪ اﻏﻠﺐ در ﻟﺤﻈﺎت اوﻟﯿﻪ ﻓﺮآﯾﻨﺪ ﻧﻔﻮذ رخ ﻣﯽدﻫﺪ، ﺑﯿﺸﺘﺮ ﺑﻮد. ﺑﻨﺎﺑﺮاﯾﻦ ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ ﻫﻤﺒﺴﺘﮕﯽ ﺑﺎﻻي ﻣﺪلﻫﺎي ﮐﻮﺳﺘﯿﺎﮐﻮف و ﻓﯿﻠﯿﭗ ﺑﺎ دادهﻫﺎي اﻧﺪازهﮔﯿﺮي ﺷﺪه در ﺣﻮزه آﺑﺨﯿﺰ ﺗﻬﻢﭼﺎي، اﯾﻦ دو ﻣﺪل ﺑﺮاي ﭘﯿﺶﺑﯿﻨﯽ ﺷﺪت ﻧﻔﻮذ در ﮐﺎرﺑﺮيﻫﺎي ﻣﺨﺘﻠﻒ زﻣﯿﻦ ﭘﯿﺸﻨﻬﺎد ﻣﯽﺷﻮد.

Accurate prediction of water infiltration rate is very important for programming the soil and water conservation practices in catchments .In this study, the efficiency of infiltration rate models was investigated in three land uses consist of dry-farming, irrigated farming and pasture lands in Tahamchai catchment (20 locations with three replications) in north west of Zanjan province, Iran. The efficiency of models was assessed using the measured data based on the determination coefficient (R2), root mean squared error (RMSE) and Nash-Sutcliffe Efficiency (NSE). Results indicated that there is obvious variation in water infiltration rate in different land uses, so that the highest value was in dry-farming lands (16.6 cm h-1) and the lowest value was in pastures (5.4 cm h-1). Among the proposed models, the Kostiakove’s equation showed the highest R2 (0.99), the lowest error (RMSE= 0.01 cm h-1) and the highest NSE (0.99) in pasture. Philip’s equation appeared to be more efficient in the rainfed-farming and irrigated farming lands with R2=0.99, RMSE ranging from 0.14 to 0.22 and NSE=0.99. Although, the Green-Ampt’s equation was fitted relatively well to the measured data (R2= 0.94), its error index (RMSE= 2.32 cm h-1) was relatively high, indicating the least efficiency among the models. The GA model showed higher estimation error for higher infiltration rate values which occur mostly at initial times of water infiltration process. Therefore, regarding the higher correlations of the Kostiakove’s and Philip’s equations with the measured data, these two models are recommended for predicting water infiltration rate in different land uses of Tahamchai catchment.