بررسي توانمندي مدل SDSMدر ريزمقياس نمايي دما و بارش در اقليم گرم و خشك (بررسي موردي: ايستگاه‌هاي همديدي يزد و طبس)

Capability assessment of SDSM model in downscaling of temperature and precipitation in hot and dry climate (case study: Synoptic stations of Yazd and Tabass)

ازآنجاكه سامانه‌هاي انساني مانند كشاورزي و صنعت، كه وابسته به عنصر‌هاي اقليمي اند، بر مبناي ثبات و پايداري اقليم طراحي شده و عمل مي‌كنند؛ ضروري است تغييرات بلندمدت دما و بارش، كه مهم‌ترين چالش در قلمرو علوم محيطي است، شناسايي شود و مدنظر قرار گيرد. براي پيش‌بيني بلندمدت عنصر‌هاي اقليمي در دوره‌هاي آتي، استفاده از مدل‌هاي اقليمي جهاني (GCMs)اجتناب‌ناپذير است. به علت تفكيك درشت ياخته محاسباتي GCMs، ضروري است براي پيش‌بيني‌هاي مقياس محلي و ناحيه‌اي از روش‌هاي ريزمقياس نمايي براي تبديل داده‌هاي بزرگ‌مقياس به داده‌هاي مقياس‌ محلي و ناحيه‌اي استفاده شود. هدف پژوهش حاضر، بررسي توانمندي مدل SDSMدر اقليم گرم و خشك براي ريزمقياس نمايي دما و بارش حاصل از خروجي مدل HadCM3تحت سناريوي A2است. در اين راستا از داده‌هاي روزانه بازتحليل NCEP/NCARو ايستگاهي دما و بارش در دوره 19612001 و داده‌هاي خروجي مدل HadCM3تحت سناريوي A2در دوره 1961-2001 شامل دما و بارش براي توليد سناريوي آتي با مختصات ايستگاه‌هاي همديدي يزد و طبس استفاده مي‌شود. مقايسه نتايج حاصل از تحليل آماري براي هر دو مجموعه داده مشاهداتي و ريزمقياس نمايي ‌شده نشان مي‌دهد كه، مدل SDSMدر ريزمقياس نمايي دماي خروجي مدل HadCM3در اقليم گرم و خشك به‌درستي عمل مي‌كند. بارش روزانه حاصل از ريزمقياس نمايي به‌كمك مدل SDSMدر اقليم گرم و خشك با داده مشاهداتي در اغلب آماره‌ها از جمله حداكثرها و حداقل‌هاي بارش تفاوت بارزي دارد. فقط برخي از آماره‌ها در مورد بارش مانند جمع ماهانه و حداكثر روزهاي خشك متوالي با داده‌هاي مشاهداتي همخواني دارند.

Since human systems such as agriculture and industry, which depend on climatic elements, are designed and created based on compatibility and stability of climate, it is essential that the longterm changes of temperature and precipitation, which constitute the most important chanllenges in the environmental sciences, are identified and considered. In ordet to longterm forecast climatic elements for future periods, the use of Global Climate Models (GCMs) is inevitable. Typically, GCMs have a resolution of 150300 km in each horizontal direction. Many impact applications require the equivalent of pointwise climate observations and are highly sensitive to finescale climate variations that are parameterized in coarsescale models. Due to the coarseresolution of the computational cell of GCMs, it is essential to use a downscaling procedure in order to convert largescale data to regional/localscales data. Downscaling aims to obtain fineresolution climate or climate change information from relatively coarseresolution GCMs. In general, downscaling is divided into dynamical and statistical categories. Dynamical downscaling fits output from GCMs into regional meteorological models such as Weather Research Forecasting (WRF) model. Thus, due to the fineresolution (2060 km) of the limited area models, it is possible to simulate some regional climatic features such as orographic precipitation, cloudiness, and some exetrem events. In climatological and meteorological researches using dynamic downscaling, a researcher can achieve both globalscale projections down to a regional/localscale and the effect of global patterns on local weather conditions. The amount of computations involved in dynamical downscaling makes it computationally expensive to produce decadeslong simulations with different GCMs or multiple emissions scenarios. The statistical downscaling method is created based on statistical relationships that link the largescale atmospheric variables with local/regional climate variables. This method has many advantages such as being easy to apply, and computationally economical. As a result, in most regional/local researches, statistical downscaling is used to consider potential impacts on specific regions or stations. In this method using appropriate statistical relationships between predictor and predictand variables, it is possible to determine the relationships for future periods. In general, if the longterm data exist for the desired station, the best method is statistical downscaling. To determine the best statistical method for downscaling in each region, it is necessary to investigate the capabilities of various statistical methods. The aim of the present research is to investigate the capability of Statistical Downscaling Model (SDSM) in a hot and dry climate to downscale temperature and precipitation as output from Hadley Centre Coupled Model, version 3 (HadCM3) under scenario A2. Several modeling tools are employed in generating the sets of Intergovernmental Panel on Climate Change (IPCC) emission scenarios. The scenario A2 is one of the IPCC emission scenarios. This scenario is based on the following assumptions; a relatively slow demographic transition and relatively slow convergence in regional fertility patterns, b relatively slow convergence in interregional gross domestic product per capita differences, c relatively slow enduse and supplyside energy efficiency improvements, d delayed development of renewable energy, and e no barriers to the use of nuclear energy. As mentioned earlier, characteristically dry and hot climate is considered to evaluate the performance of SDSM model. Therefore, daily NCEP/NCAR reanalysis and station data during the 19612001 period and output from HadCM3 under scenario A2 for 19612001 period containing temperature and precipitation for Yazd and Tabas synoptic stations are used. Comparing the results obtained from statistical analyses for observational and downscaled data indicates that the SDSM model can downscale correctly temperature output from HadCM3 in hot and dry climates. Daily precipitation resulted from downscaling using SDSM model has marked differences with observational precipitation in most of the statistical quantities used such as maximum and minimum precipitation. Only some statistical quantities such as the sum of the monthly precipitation and maximum consecutive dry days are consistent with the observed data.