ارزيابي مدل هاي جبري و زمين آماري در تخمين توزيع مكاني سطح ايستابي دشت اردبيل

Deterministic models and statistical evaluation of the estimated spatial distribution of Ardabil Plain water table

پژوهش حاضر به منظور مقايسه مدل هاي جبري و زمين آماري در پهنه بندي عمق آب‌هاي زيرزميني دشت اردبيل در يك بازه زماني 28 ساله( 1362 تا 1390)، با استفاده از داده‌هاي 22 حلقه چاه پيزومتري در سطح دشت اردبيل به انجام رسيده است. مدل هاي مقايسه شده مدل IDW و مجموعه روش‌هاي توابع شعاعي(RBF)، به عنوان نماينده روش‌هاي جبري و مدل كريجينگ به عنوان نماينده روش‌هاي زمين آماري مي‌باشد. از اعتبارسنجي متقاطع، شاخص هاي خطاي، ميانگين انحراف خطا(MBE)، ريشه دوم ميانگين مربع خطا(RMSE)، همبستگي مقادير برآوردي و مشاهده‌اي، واريانس تخمين جهت انتخاب روش مناسب و اعتبارسنجي مدل هاي مربوطه استفاده‌شده است. مدل انتخاب شده با نقشه كاربري اراضي منطقه، پراكندگي چاه‌هاي عميق و نيمه عميق، همچنين شبكه آبراهه‌هاي سطح دشت مورد مقايسه قرار گرفت تا از صحت و دقت برآورد صورت گرفته اطمينان حاصل گردد. نتايج نشان داد از بين مدل هاي مختلف مدل Completely Regularized spline كه يكي از مدل هاي توابع شعاعي محسوب مي‌شود نسبت به ساير روش‌ها از دقت بالاتري برخوردار بوده و با نقشه‌هاي كاربري اراضي منطقه، پراكندگي چاه‌هاي عميق و نيمه عميق و شبكه آبراهه‌هاي منطقه همخواني بسيار بالايي دارد. بنابراين بهترين روش جهت پهنه بندي عمق آب‌هاي زيرزميني دشت اردبيل محسوب مي‌گردد.

Extended Abstract Introduction Groundwater from the Earthʹs fresh water supply is one of the largest glaciers. Groundwater plays a key role in productive sectors such as agriculture and animal husbandry Iran. Due to Iran standing in hot and dry region, Very limited amount of precipitation received and doesn’t provide the daily needs of the industrial sector, agriculture and animal husbandry, that’s why the main reason for the countryʹs water source are supplied from groundwater. Ardebil plain is not in exception of being fertile and having the best potential for agriculture and livestock farming of the land of the plain form, Due to insufficient precipitation for these activities, Abundant in shallow to deep wells in the desert, has been lead to a sharp drop in groundwater levels in some sectors; And is considered as one of the forbidden fields of deep wells in the deep half of the country. So the precise knowledge of the groundwater level anywhere in the vast plains of Ardabil Because of possible measures to prevent the indiscriminate harvesting of groundwater in plain area is essential. Geostatistical model is used in order to identify one of these well-known factors around the world and always point to generalize the data to estimate the unknown quantities. Methodology Methods used in this study are algebraic models and geostatistics that their description is as follows: Algebraic methods: Algebraic approach procedures at one or more points of view to a (z) are fitted to the given coordination. Algebraic interpolation can, so they are approximative or an exact amount so if the observed value as exact values (no error or uncertainty) in the sample is taken, it is recommended to use a method for accurate interpolation. However, some uncertainty may be considered for a smoothly varying selection. Therefore, these different functions can be used to fit the level of interpolation points. Geostatistical methods: Within statistical pitch detection method that is based on a set of statistical features a series of spatial variables, such as mean and standard deviation. Statistical methods to estimate the spatial average of a region depends on the number of samples and the possible role of each sample. Therefore we can say they are doing best estimation of the statistical methods and the validity of some of the estimates is calculated. The model is also used to assess the cross-validation, error indicators, Mean Bias Error (MBE), Root mean square error (RMSE), Correlation between the observed and estimated values, variance estimation and appropriate method used for validating the models. Result and discussion: The results of different methods for finding the middle of Ardabil Plain interpolation water table are as follows: Method of inverse distance weighted IDW: The best we can for the zoning of groundwater depth of at least 2 and up to 15 fields of Ardabil 1 and neighboring areas were used for zoning. The error in the estimation of this RMSE and ME Ardabil Plain groundwater depth is 9/28 and 0.24 respectively. RBF techniques are radial functions: The method of the radial functions of the five models in this family Spline covers were used; that Completely Regularized spline model among other methods of estimation covers a minimal error, therefore, this method was used for the study. This method of estimation error RMSE and ME is 7/99 and 0.71 respectively. Kriging Method: Necessary condition for using this method is data normality, or at least closing normal distribution of data. Because of the abnormal distribution of data we used the Log Normal distribution method in order to normalize the distribution. The semivariogram without direction to semivariogram different direction, and the theoretical models that best fit the semivariogram Hole Effect was used for this procedure. ME and RMSE error of the method in Ardabil Plain groundwater depth zonation is 8/71 and 1/53 respectively. Conclusion: The results of the various models shows that, the model Completely Regularized spline that is one of the radial functions models; in comparison to other methods has higher accuracy, and the land use map of the distribution of deep and semi-deep wells and a network of regional channels are in high association. So the best method for zoning of Ardabil Plain groundwater depth is considered. The analysis suggests a dangerous groundwater situation in the southeastern part of Ardabil plain; due to excessive concentration in shallow and deep wells in the area and picking up a lot of its power; cause a severe drop in groundwater levels in this area that could result in very bad consequences, such as severe land subsidence and groundwater aquifers drying up in these areas to be followed; so is as a dangerous disaster in Ardabil plain, thatʹs why you should check the indiscriminate digging of wells and using groundwaterʹs, necessary measures should be taken, so this sectors to the status of groundwater level has improved and groundwater levels in these areas with better nutrition and taking appropriate to be distributed in lower depth to the surface.