معرفی و ارزیابی مدلی پایدار برای تخمین شاخص سطح برگ گندم بوسیله تصاویر ماهواره‌ای در شرایط اقلیمی متفاوت

تخمین دقیق شاخص سطح برگ به عنوان یکی از متغیرهای کلیدی اکوسیستم دارای اهمیت فراوانی می‌باشد. هدف از این تحقیق معرفی مدلی مناسب جهت برآورد شاخص سطح برگ گندم بوسیله ndvi محاسبه شده از تصاویر لندست و مقایسه دقت آن با مدل‌های رایج آماری می‌باشد. در این راستا، داده‌های چندساله در دو منطقه در جنوب غرب آلمان برای واسنجی و اعتبارسنجی مدل‌های تجربی، نیمه تجربی و مدل معرفی شده در این مقاله با نام nple مورد استفاده قرار گرفت. سپس مقادیر بهینه پارامترهای هر مدل برای ارزیابی در مزارع آستان قدس مشهد استفاده شد. مدل‌های مورد ارزیابی عبارتند از: 1 وینا (مدل تجربی سه پارامتری). 2 لیو (مدل نیمه‌تجربی با دو پارامتر تجربی و یک پارامتر مربوط به ضریب خاموشی). 3 چادوری (مدل نیمه تجربی با پارامتر ضریب خاموشی) و 4 nple (مدل تجربی بدون پارامتر) و 5 نسخه اصلاح شده‌ی nple که در آن بجای ndvi از درصد پوشش گیاهی استفاده شده است. نتایج نشان داد که مدل‌های nple و وینا دارای خطای بیشتری نسبت به بقیه مدل‌ها بودند. مدل لیو، چادوری و nple اصلاح شده دارای خطای قابل قبولی در واسنجی و اعتبارسنجی بودند (rmse~0.30). در مرحله ارزیابی مدل‌ها در مشهد، مدل nple اصلاح شده و چادوری بهترین نتیجه را داشتند، در حالی که مدل‌های دیگر دارای خطای سیستماتیک بالایی بودند، این مطلب نشان می‌دهد که مقادیر بهینه بدست آمده برای پارامترهای مدل‌های لیو و وینا با تغییر مکان آزمایش (مشهد) معتبر نمی‌باشند. بدلیل کمتر بودن و همچنین غیر سیستماتیک بودن خطای پیش بینی در مدل nple اصلاح شده، پیشنهاد می‌شود که خروجی این مدل در مطالعات واسنجی مدل‌های زراعی مورد استفاده قرار گیرد

Introducing and evaluating a robust statistical model to estimate wheat LAI from satellite data in different climatological locations

IntroductionOne of the most interesting research areas of the scientific community in recent years is the study of the behavior and regimes of rainfall at the local, regional and global levels. At the proposal of the Intergovernmental Panel on Climate Change, rainfall changes should be comprehensively addressed in all areas (IPCC, 2007). The zoning and recognition of homogeneous climatic regions is one of the basic needs of planning. Climatic zoning is often based on the use of different climate variables in order to take into account the role of all variables in determining the climate of the regions. The main feature of rainfall in Iran is that the annual precipitation in the country is significant both in spatial and temporal terms. This spatial and temporal distribution of rainfall in Iran is affected by the distribution of global circulation systems, which has the slightest change in its pattern, causing extreme weather abnormalities. Therefore, spatial and temporal abnormalities of rainfall and severe changes in rainfall intensity and rainfall difference are the most important characteristics of rainfall in Iran (Babaei and Farajzadeh, 2002: 52). The temporal and spatial distribution of rainfall in Azerbaijan is also important due to its impact on Iranian water resources. In this study, it has been tried to analyze the cluster statistical methods to identify the zoning and zoning regime in Azerbaijan. To be Cluster analysis is a statistical method that clustering a set of individuals in terms of their similarity. Therefore, each cluster is a group whose constituents have the most similarity. There have been a lot of rainfall related to the parameters that can be mentioned.Masoudian (2005) determined the cluster analysis method for Iran rainfall and concluded that Iran has three main regimes: first winter precipitation regime, second winter / spring precipitation regime and third winter weather regime. Mohammadi (2011) investigated Iranian precipitation trend using data from 1437 stations of synoptic, climatic and barometric stations during a 40year period and concluded that in the time series of Iran’s precipitation and pixel average, there was a significant increase or decrease in the confidence level is 95 percent (Mohammadi, 2011). In an exploratory study, Younesi (2014) explores the climate of the Republic of Azerbaijan and concludes that the maximum rainfall of Azerbaijan occurs in the northern parts (highlands of Shah dagh and Bazarzdozo), southeast (Lankaran region) and western (Khan Kandi area). Masterzard (2014) study of the terrestrial climate in Afghanistan based on the daily rainfall data of the Aphrodite database. The results of this study have shown that the average rainfall in Afghanistan is 256 mm. The area has a maximum of two parts in the eastern region with 800 mm in terms of rainfall, and another in the northeast with 450 mm. Mefakheri et al. (2017) investigated the time and uniformity of precipitation in Iran. In this research, cluster analysis method for climatic zoning and to estimate the spatial and temporal dispersion of rainfall and spatial data, the coefficient of variation and uniformity of statistics have been used in three decades, years, and seasonal periods. Van et al. (2011) conducted a research entitled Spacecraft spatial trends in the Les Cheson plateau. The results showed that based on the nonparametric MannKendall test, there are no significant changes in the precipitation trend of this region. Theodoro et al. (2016) studied the temporal and spatial variations of monthly rainfall in the Brazilian Mato Grosso region using a cluster analysis. Rao et al. (2017) evaluated the precipitation trend in the central and southern regions of Peru during the statistical period from 1965 to 2010. Using cluster and component analysis, they identified four areas of rainfall and examined the trend of rainfall changes in these areas.The country of Azerbaijan is important due to its neighbors with our country and because of its historical, cultural, ethnic and natural ties with Iran. Identifying its climate features is also one of the areas that can be of interest to Iranian researchers. This research has been done for this reason.Materials and methodsThe South Caucasus region includes the countries of Azerbaijan, Armenia and Georgia. Azerbaijan is located in the south of the Caucasus Mountains and north of the Aras River, near the Caspian Sea. In this study, the daily rainfall data of the Aphrodite database, which has a spatial resolution of 0.25 * 0.25 degrees, has been used in a statistical period of 60 years (19511951) to identify the rainfall regions of Azerbaijan. At first, rainfall data of 263 locations (pixels) located in the territory of Azerbaijan and adjacent regions were prepared from the base of Aphrodite, and then the average daily precipitation was calculated for these points and a 365 × 263 matrix was formed. This matrix was the basis for judging the peripheral regions and perihelion regimes of Azerbaijan. In order to identify the terrestrial and rainy regions of the country, the cluster analysis of the basin was used by integrating into MATLAB software.Results and discussionThe average annual rainfall of Azerbaijan is 397 mm. The maximum rainfall of Azerbaijan is observed at the northern altitudes of Shah Dagh and Bazarduzu, and lowlying areas in Lankaran with values 600 mm and a minimum rainfall in the central regions of Shirvan and Baku with values  300 mm. As the north and west of the region progress to the central and southern parts, the annual precipitation is also reduced. In Nakhchivan, the maximum rainfall of about 450 mm corresponds to the mountainous areas of the northeast and as far as the south and west of the Nakhchivan area advance, annual rainfall decreases.Precipitation is a climate whose amount changes continuously. By analyzing the cluster on the average annual rainfall in Azerbaijan, this country is divided into six districts. These areas are different in terms of precipitation and annual distribution. In general, the country of Azerbaijan is divided into two high rainfall (the Lankaran and high Caucasian) and low rainfall (coastal area and Baku). However, each of the high and low rainfall regions was divided into smaller areas and finally six areas were identified in Azerbaijan.In order to identify the regime of the country of Azerbaijan, a matrix with a size of 363 × 363 was formed, which represented 263 places per day in terms of percentage per day. On this basis, the first Euclidean distance of all spatial precipitation is measured every day. After measuring the Euclidean distance, a cluster analysis was performed by integration method on the interval matrix and 263 points were clustered in accordance with the degree of similarity. In general, the diet regime in Azerbaijan is divided into two types of rainy regimes: spring rainfall regime, which includes three subdiet regimes (Nakhchivan regime, Northwest regime and central regime), and the most annual rainfall occurs during the spring and the autumn regime Spring with two regimes (Lankaran regime and East Azarbaijan and Baku regime), in which the share of precipitation falls from the first year of precipitation and th