ارزیابی مکفی بودن بارش‌های موثر در زراعت دیم به‌منظور توسعه کشت غلات پاییزه در اراضی شهرستان های آق قلا، کلاله و گنبدکاووس (استان گلستان) با استفاده از تحلیل‌های مکانی gis

در استان گلستان با توجه به وجود دیم‌زارهای گسترده، مقدار و توزیع مکانی و زمانی بارش، از اهمیت زیادی برخوردار است. در این مطالعه، با استفاده از تحلیل‌های مکانی سامانه اطلاعات جغرافیایی، تغییرات مکانی و زمانی و نیز کفایت متغیر های بارش در محدوده اراضی کشاورزی شهرستان های آق قلا، کلاله و گنبدکاووس جهت توسعه زراعت دیم غلات پاییزه مورد ارزیابی قرار گرفت. این پژوهش در سال 1394 در دانشگاه علوم کشاورزی و منابع طبیعی گرگان اجرا گردید. ابتدا نقشه های موضوعی انواع متغیرهای بارش موثر در زراعت دیم شامل بارش سالانه، بارش پاییزه، بارش بهاره، بارش آبان ماه و بارش اردیبهشت ماه، با استفاده از روش های متعدد درون یابی مانند کریجینگ معمولی، وزن دهی فاصله معکوس (idw) و توابع پایه شعاعی (rbf) در محیط gis نسخه 10 تهیه و طبقه بندی شدند. از معیارهای میانگین مطلق خطا (mae)، میانگین انحراف خطا (mbe) و خطای برآورد ((rmse، جهت ارزیابی این روش ها استفاده شد. در این پژوهش از داده های 10 ایستگاه هم دیدی و 77 ایستگاه باران سنجی مستقر در استان گلستان استفاده گردید. نتایج نشان داد که روش کریجینگ با مدل کروی بهترین الگو برای تخمین بارش پاییزه و اردیبهشت ماه و هم چنین روش توابع پایه شعاعی مدل چند ربعی، بهترین الگو برای تخمین بارش های سالانه، بهاره و آبان ماه در اراضی کشاورزی شهرستان های آق قلا، کلاله و گنبدکاووس می باشد. روش وزن‌دهی فاصله معکوس توان 1 و روش توابع پایه شعاعی مدل نواری کم ضخامت، نامناسب ترین الگو جهت تخمین انواع بارش شناخته شد. نتایج توزیع مکانی بارش نشان داد از سمت شمال به جنوب مناطق موردمطالعه، بر میزان متغیرهای بارش افزوده می شود. بر این اساس، برخی از مناطق شمالی شهرستان های آق قلا و گنبدکاووس ازنظر بارش بهاره و بارش اردیبهشت ماه جهت توسعه زراعت دیم دارای محدودیت می باشند. درمجموع در بین سه شهرستان موردبررسی ، شهرستان کلاله بیش ترین میزان و بهترین توزیع مکانی انواع متغیرهای بارش را جهت توسعه زراعت دیم دارا بود.

Investigation of effective precipitation sufficiency in rainfed farming in order to development of autumn cereals cropping in agricultural lands of AqQala, Kalale and Gonbad Kavous townships (Golestan province) by spatial analysis of GIS

Introduction: ;Annual climate variation, along with water shortage is an important factor influencing agricultural productivity, especially in rainfed crop production systems. Also, understanding the current relationship between food grain production and rainfall will help in assessing the possible impact of future climate change on the livelihood of smallholder farmers. In Golestan province due to different climates and extended rainfed lands, amount and spatial and temporal distribution of rainfall is very important. In general, the minimum rainfall for dry farming is about 250300 mm. In addition, distribution of rainfall during the growing period is also noticeable. The enough moisture in vegetation phase is essential for growth and establishment of plant. In this study, spatial analysis of GIS and some methods of classic interpolation and geostatistical methods were used to evaluate spatial and temporal changes of rainfall variables in the agricultural lands of AqQala, Kalale and Gonbad Kavous townships in Golestan province.;  ;Materials and methods:;This research was carried out in Gorgan University of Agricultural Sciences and Natural Resources (GUASRN), during 2015. The studied region was included the agricultural lands of AqQala, Gonbad Kavous and Kalale townships in Golestan province, as three important regions for dry farming production. First, the data of 10 synoptic stations and 77 raingauge stations from Golestan province were used. The thematic maps of rainfall variables such as annual, autumn, spring, November and May precipitations were provided using various methods of geostatistical and classic interpolation such as Ordinary Kriging (OK), Inverse Distance Weighting (IDW) and Radial Basis Function (RBF) in GIS (var. 10) media. Before interpolation calculations, common statistical tests were performed on main data. The performance criteria for evaluation were Mean Absolute Error (MAE), Mean Bias Error (MBE), Root Mean Square Error (RMSE) and Cross Validation. In this research, a semivariogram function was used to show the variation of rainfalls with respect to distance.;  ;Results and discussion: ;The results showed that Krigingspherical model was the best method to estimate autumn and May precipitations and also, the annual, spring and November precipitations were interpolated by Multiquadric (RBF) in agricultural lands of AqQala, Kalale and Gonbad Kavous townships. Furthermore, the IDW and Thin Plate Spline (TPS) were introduced as inappropriate methods due to high error in precipitation estimate. The semivariograms analysis indicated that autumn, spring, November and May precipitation variables were the best fitted by Spherical models, but annual precipitation was fitted by Exponential model. It was found that some areas in north of AqQala and Gonbad Kavous had the limitation in amounts of spring and May precipitations. The spatial distribution of precipitations showed that annual precipitation was increased from north to south, so that, the maximum annual precipitation was observed in Kalale township. Thus, this township can be a region with high suitable degree for dry farming some autumn crops such as wheat and barley. The lowest autumn rainfall amount was observed in the northern parts of Gonbad Kavous township with 31.37 mm, but the southern parts of the AqQala, Gonbad Kavous and especially Kalale had the highest autumn precipitation amounts.;  ;Conclusion:;Result of the spatial distribution of precipitation showed that amounts of precipitation variables were increased from north to south of the studied region. Totally, among the three studied townships, Kalale township had the highest amounts of precipitations and the best spatial distribution of precipitation variables for rainfed farming. Therefore, this township was selected as the high suitable region for dry farming in Golestan province.;  Acknowledgements: ;We are grateful to Agriculture Service Centers of Golestan province and Gorgan University of Agricultural Sciences and Natural Resources (GUASNR).