پديد آورندگان :
ﭘﺮﭼﻤﯽ ﻋﺮاﻗﯽ، ﻓﺮزﯾﻦ سازمان تحقيقات، آموزش و ترويج - مركز تحقيقات و آموزش كشاورزي و منابع طبيعي استان اردبيل (مغان) - بخش تحقيقات فني و مهندسي كشاورزي , ﺻﺎدﻗﯽ ﻻري، ﻋﺪﻧﺎن دانشگاه هرمزگان - دانشكده كشاورزي - گروه مهندسي آب
كليدواژه :
بهينه‌سازي رفتار جمعي اجزاي يكپارچه , شوري , مدل SWAP , مدل‌سازي معكوس
چكيده فارسي :
ﺷﺒﯿﻪ ﺳﺎزي ﻫﺎي زراﻋﯽ-ﻫﯿﺪروﻟﻮژﯾﮑﯽ در ﻣﻌﺮض درﺟﺎت ﻣﺨﺘﻠﻔﯽ از ﻋﺪم ﻗﻄﻌﯿﺖ ﻗﺮار دارد. در اﯾﻦ راﺑﻄﻪ، ﺗﺤﻠﯿﻞ ﻋﺪم ﻗﻄﻌﯿﺖ ﻣﯽ ﺗﻮاﻧﺪ ﺑﯿﻨﺶ ﻣﻔﯿـﺪي
در ﺧﺼﻮص درﺟﻪ اﺳﺘﺤﮑﺎم ﺧﺮوﺟﯽ ﻫﺎي ﻣﺪل ﻓﺮاﻫﻢ آورد. در اﯾﻦ ﻣﻄﺎﻟﻌﻪ، ﺑﺎ ﺗﻠﻔﯿﻖ روش ﺑﺮآورد ﻋﺪم ﻗﻄﻌﯿﺖ درﺳﺖ ﻧﻤﺎﯾﯽ ﺗﻌﻤﯿﻢ ﯾﺎﻓﺘﻪ )GLUE( و ﮔﻮﻧﻪ ﯾﮑﭙﺎرﭼﻪ ﺳﺎزي ﺷﺪه اﻟﮕﻮرﯾﺘﻢ ﺑﻬﯿﻨﻪﺳﺎزي رﻓﺘﺎر ﺟﻤﻌﯽ اﺟﺰا )UPSO( ﯾﮏ ﻃﺮح ﺗﺤﻠﯿﻞ ﻋﺪم ﻗﻄﻌﯿـﺖ ﭼﻨﺪﻫﺪﻓـﻪ ﺗﺮﮐﯿﺒـﯽ )GLUE-UPSO( ﻣﺒﺘﻨـﯽ ﺑـﺮ ﺷﺒﯿﻪ ﺳﺎزي ﻫﺎي ﻣﻬﺎر ﺷﺪه ﻣﻮﻧﺖ ﮐﺎرﻟﻮ ﺗﻮﺳﻌﻪ ﯾﺎﻓﺖ. ﻃﺮح ﺗﻮﺳﻌﻪ ﯾﺎﻓﺘﻪ ﺑﻪ ﻣﻨﻈﻮر ﺗﺤﻠﯿﻞ ﻋﺪم ﻗﻄﻌﯿﺖ ﻣﺪل ﺳﺎزي زراﻋﯽ-ﻫﯿﺪروﻟﻮژﯾﮑﯽ ﺗﻮزﯾﻌﯽ ﺑﺎ اﺳـﺘﻔﺎده از ﻣﺪل SWAP ﺑﺮاي ﯾﮏ ﻣﺰرﻋﻪ ﻧﯿﺸﮑﺮ ﺑﺎ زﻫﮑﺸﯽ زﯾﺮزﻣﯿﻨﯽ، واﻗﻊ در ﮐﺸﺖ و ﺻﻨﻌﺖ ﻧﯿﺸﮑﺮ ﺷﻌﯿﺒﯿﻪ، ﺧﻮزﺳﺘﺎن ﻣﻮرد ﮐﺎرﺑﺮد ﻗﺮار ﮔﺮﻓﺖ. ﻧﺘﺎﯾﺞ اﯾﻦ ﺑﺮرﺳﯽ ﺣﺎﮐﯽ از ﻏﯿﺮﯾﮑﺘﺎﯾﯽ ﻗﻮي اﮐﺜﺮ ﭘﺎراﻣﺘﺮﻫﺎي واﺳﻨﺠﯽ ﺷﺪه و اﻫﻤﯿﺖ ﺗﺤﻠﯿﻞ ﻋﺪم ﻗﻄﻌﯿﺖ ﺷﺒﯿﻪ ﺳﺎزي ﻫﺎي ﻣﺪل SWAP ﺑﻮد. وﺟﻮد ﺿﺮاﯾﺐ ﻫﻤﺒﺴﺘﮕﯽ ﻗﻮي ﺑﯿﻦ ﭘﺎراﻣﺘﺮﻫﺎي واﺳﻨﺠﯽ ﺷﺪه ﺣﺎﮐﯽ از اﻫﻤﯿﺖ واﺳﻨﺠﯽ ﻣﻮﻟﻔﻪ ﻫﺎي ﻣﺨﺘﻠﻒ ﻣﺪل ﺑﻪ ازاي داده ﻫﺎي اﻧﺪازه ﮔﯿﺮي ﺷﺪه ﻣﻌﯿﺎر ﻣﺘﻨﻮع ﺑﻪ ﺻـﻮرت ﻫـﻢ زﻣـﺎن ﺑـﻮد. ﻣﺤﺪوده ﻫﺎي ﻋﺪم ﻗﻄﻌﯿﺖ ﭘﯿﺶ ﺑﯿﻨﯽ 95 درﺻﺪ اﺷﺘﻘﺎق ﯾﺎﻓﺘﻪ ﺑﺮاي ﻣﻮﻟﻔﻪ ﻫﺎي ﻫﯿﺪروﻟﻮژي )رﻃﻮﺑﺖ ﺧﺎك، ﻧﻮﺳﺎﻧﺎت ﺳﻄﺢ اﯾﺴﺘﺎﺑﯽ و ﺟﺮﯾﺎن زه آب ﺧﺮوﺟﯽ از زﻫﮑﺶ زﯾﺮزﻣﯿﻨﯽ(، اﻧﺘﻘﺎل اﻣﻼح )ﻧﯿﻢ رخ ﻏﻠﻈﺖ اﻣﻼح آب ﺧﺎك و ﺷﻮري زه آب( و ﺑﯿﻮﻓﯿﺰﯾﮑﯽ ﻣﺪل )ﺷﺎﺧﺺ ﺳﻄﺢ ﺑﺮگ، ﻋﻤﻠﮑﺮد ﻧﯽ و ﻋﻤﻠﮑـﺮد ﺳـﺎﮐﺎرز( به‌ترتيب، بين 67 تا 90، 23 تا 71 و 75 تا 100 درصد از كل داده‌هاي اندازه‌گيري شده (اعم از هردو مجموعه داده‌هاي واسنجي و صحت‌سنجي) را با r-factor بين 0/58 ﺗﺎ 0/43 ،1/34 ﺗﺎ 1/07 و 0/70 ﺗﺎ 0/98 در ﺑﺮ ﮔﺮﻓﺘﻨﺪ. ﻧﺘﺎﯾﺞ اﯾﻦ ﻣﻄﺎﻟﻌﻪ، ﺣﺎﮐﯽ از ﺳﻄﺢ ﻗﺎﺑﻞ ﻗﺒﻮل ﻋﺪم ﻗﻄﻌﯿﺖ ﻣﺪل و ﻗﺎﺑﻠﯿـ ﺖ روش ﺗﻠﻔﯿﻘﯽ ﺗﻮﺳﻌﻪ ﯾﺎﻓﺘﻪ در واﺳﻨﺠﯽ و ﮐﻤﯽ ﺳﺎزي ﻋﺪم ﻗﻄﻌﯿﺖ ﻣﻮﻟﻔﻪ ﻫﺎي ﻣﺨﺘﻠﻒ ﻣﺪل ﺑﻪ ﻃﻮر ﻫﻢ زﻣﺎن ﺑﻮ
چكيده لاتين :
The agro-hydrologic simulations are subject to varying degrees of uncertainty. In this regard, uncertainty analysis can provide useful insights into the level of robustness of the model outputs. In this study, a hybrid multi-objective uncertainty analysis scheme, GLUE-UPSO, based on strained Monte-Carlo simulations was developed, combining Generalized Likelihood Uncertainty Estimation (GLUE) and Unified Particle Swarm Optimization (UPSO). The developed scheme was used for uncertainty analysis of distributed agro-hydrological modeling with SWAP for a sugarcane farming system with subsurface drainage located at Shoaybiyeh Sugarcane Agro-industrial Company farms, Khuzestan, Iran. The results highlighted the strong nonuniqueness of most of the calibrated parameters and the importance of uncertainty analysis of the SWAP simulations. Strong parameter correlations revealed the need for the simultaneous calibration of the model parameters against diverse calibration data. The 95% prediction uncertainty bands obtained for the model's hydrology (soil water content, water table level, sub-surface drainage outflow), solute transport (soil water solute concentration and sub-surface drainage outflow salinity), and biophysical (leaf area index, cane, and sucrose dry yield) components enveloped 67%-90%, 23%-71%, and 75%-100% of the corresponding observed data (including both calibration and validation datasets), respectively, with an r-factor of 0.58-1.34, 0.43-1.07, and 0.70-0.98. The results of the study indicated the acceptable level of model uncertainty and the capability of the developed framework for simultaneous calibration and uncertainty quantification of model components.
