CAJ | 학술논문

参数敏感性分析是模型不确定性量化的重要环节，有助于对关键参数的识别，减少参数的不确定性影响，进而提高参数优化效率。以太湖地区典型小流域为研究区，采用扰动分析法对 HSPF 模型水文模块、泥沙模块以及氮磷输移等水文、水质模拟过程的参数进行了敏感性分析。研究结果显示：水文模块选取的17个参数中有7个敏感：UZSN、INFILT、AGWRC 对径流的敏感级别为芋类，LZSN、DEEPFR、INTFW、IRC 敏感级别为域类。泥沙透水地面模块选取的9个参数中，KSER、KGER、JGER 为芋类敏感参数， JSER 为郁类敏感参数；不透水地面模块选取的4个参数中，KEIM、JEIM、ACCSDP 对泥沙产量的敏感级别为芋类；河道模块选取的5个参数中，KSAND、EXPSND 为芋类敏感参数，TAUCS、TAUCD 为域类敏感参数。总氮模拟选取了23个参数分析敏感性，其中WSQOP、SQOLIM、MON-GRND-CONC 为郁类敏感参数，KATM20、MON-IFLW-CONC 为芋类敏感参数，TCNIT、PHYSET、MALGR敏感级别为域类。磷素输移模拟选取了12个参数，MON-GRND-CONC 敏感级别为芋类，MON-POTFW、MON-IFLW-CONC、MALGR、PHYSET 敏感级别为域类。研究结果对于开展基于 HSPF 模型的流域水文水质研究工作参数的选取具有一定的参考价值，尤其对于太湖周边地区众多低山丘陵小流域进行 HSPF 模型水文水质模拟时敏感性参数的选取具有借鉴意义。
삼수민감성분석시모형불학정성양화적중요배절，유조우대관건삼수적식별，감소삼수적불학정성영향，진이제고삼수우화효솔。이태호지구전형소류역위연구구，채용우동분석법대 HSPF 모형수문모괴、니사모괴이급담린수이등수문、수질모의과정적삼수진행료민감성분석。연구결과현시：수문모괴선취적17개삼수중유7개민감：UZSN、INFILT、AGWRC 대경류적민감급별위우류，LZSN、DEEPFR、INTFW、IRC 민감급별위역류。니사투수지면모괴선취적9개삼수중，KSER、KGER、JGER 위우류민감삼수， JSER 위욱류민감삼수；불투수지면모괴선취적4개삼수중，KEIM、JEIM、ACCSDP 대니사산량적민감급별위우류；하도모괴선취적5개삼수중，KSAND、EXPSND 위우류민감삼수，TAUCS、TAUCD 위역류민감삼수。총담모의선취료23개삼수분석민감성，기중WSQOP、SQOLIM、MON-GRND-CONC 위욱류민감삼수，KATM20、MON-IFLW-CONC 위우류민감삼수，TCNIT、PHYSET、MALGR민감급별위역류。린소수이모의선취료12개삼수，MON-GRND-CONC 민감급별위우류，MON-POTFW、MON-IFLW-CONC、MALGR、PHYSET 민감급별위역류。연구결과대우개전기우 HSPF 모형적류역수문수질연구공작삼수적선취구유일정적삼고개치，우기대우태호주변지구음다저산구릉소류역진행 HSPF 모형수문수질모의시민감성삼수적선취구유차감의의。
Model sensitivity analysis measures the variability of output variables caused by perturbations in parameter values and input data. It is important for parameter selection, model calibration, and model improvement. As one of the integrated watershed model, HSPF(Hydro-logical Simulation Program-Fortran)model has a lot of parameters related to the physical characteristics of local watershed. In order to as-certain the sensitive parameters for hydrology and water quality simulation of HSPF model, a typical small watershed in the Taihu Lake hilly region was used as a case study. Perturbation method was employed to evaluate the sensitivity of parameters of hydrology, sediment and N and P transportation modules. It was found that seven of 17 selected parameters in the hydrology module were sensitive. The sensitive level of UZSN, INFILT, and AGWRC was Type Ⅲ, while that of LZSN, DEEPFR, INTFW, and IRC Type II. Of nine parameters for the pervious surface of sediment, the sensitive level of KSER, KGER and JGER was Type Ⅲ, and JSER Type ≤. Among the four parameters for impervi-ous surface, the sensitive level of KEIM, JEIM and ACCSDP was Type Ⅲ. The reach parameters KSAND and EXPSND had Type Ⅲ sensi-tivity, while TAUCS and TAUCD had Type II sensitivity. In selected 23 parameters for nitrogen simulation module, WSQOP, SQOLIM, MON-GRND-CONC were Type Ⅳ, KATM20 and MON-IFLW-CONC Type ⅢⅣ, and TCNIT, PHYSET and MALGR Type II sensitive. Phosphorus module had 12 parameters, of which MON-GRND-CONC was Type Ⅲ and MON-POTFW, MON-IFLW-CONC, MALGR, and PHYSET Type II sensitive parameters. These results would be useful for selection of sensitive parameters for hydrology and water quality simulation used for HSPF model, especially for small hilly watersheds in the Taihu Lake area.