بررسی تغییرات تراز دریا در اثر پارامترهای اقلیمی با استفاده از الگوریتم درخت تصمیم در سواحل شمالی دریای عمان (مکران ساحلی)

برخی عناصر اقلیمی، مانند دمای هوا، فشار هوا، و سرعت باد، در زمره شدیدترین تغییرات کوتاه‏مدت تراز سطح آب قرار می‏گیرد. در این پژوهش، تغییرات تراز آب بر پایه داده‏های ایستگاه هواشناسی و ایستگاه ثبت جزرومد با هدف ارزیابی کارآیی از مدل درخت تصمیم‏گیری (آنالیز غیرخطی) در برآورد و چگونگی اثرهای پارامترهای مستقل بارومتری، نیروی باد، و دمای هوا بر پیش‏بینی روند پارامتر وابسته میانگین تراز دریا (msl) سواحل شمالی دریای عمان (مکران ساحلی جنوب‏ شرق ایران) در ایستگاه‏های جزرومدی مناطق جاسک و چابهار در یک دوره  بیست‏ساله (19972016) ارزیابی و محاسبه شده است. براساس رابطه مدل نهایی در منطقه جاسک ((71/0)w 208/7 + (195/0) p 092/11 – (102/1) t 619/5+ 197/13msl=) و در منطقه چابهار ((316/0) w776/2+(87/0) p596/1(089/1) t529/1+520/4msl=) حاصل از الگوریتم درخت تصمیم‏گیری در پیش‏بینی msl با استفاده از داده‏های موجود تا 95درصد قابل اطمینان است. با توجه به دامنه نوسان جزرومد، به طور متوسط، در منطقه چابهار 1 تا 5/1 متر و در منطقه جاسک بیش از 3 متر برابر بازه زمانی به‏دست‏آمده در نمودارها با بررسی اجمالی اشکال ژئومورفولوژیکی در منطقه مطالعاتی از لحاظ آثار مورفولوژیک و داغ آب نیز به‏خوبی مطابقت می‏نماید.

Investigation of sea level changes due to climate parameters using decision tree algorithm On the northern coasts of Oman Sea (Makran coastal)

Extended Abstract Introduction The relationship between form and process in geomorphology is very important. By changing the process, the forms will be changed and new processes will be created in the form of new forms (Gurabi & Emami, 2017. p:75). Sea level changes mainly include tidal variations and changes due to atmospheric factors. Tidal flows are also affected by coastal washing during their daily advancement and retreat on tidal slopes and tidal zones. The formation of many coastal geomorphologic forms is the result of their involvement. (Nohegar & Hosseinzadeh, 2011. P:131). Climatic factors cause short time fluctuations and tidal cycles and long term fluctuations in medium level of sea (Yamani & MohammadNejad, 2012: p: 87). Torabi Azad and Honarmand (2016) performed a concise investigation about sea level changes in Banda Abbas and Booshehr Stations in a period of 11 years (2000 to 2010) and analyzed and computed barometric effects, wind force and temperature on the sea level mean. The results showed that sea level mean in these stations has incremental trend by 5 cm and 4 cm respectively in the mentioned seaports. Akbari et al (2017) In order to investigate and analyze important tidal components in a vast area including Persian Gulf, Hormuz Strait, Oman sea and Arab sea applied 3D FVCOM Model. The results of this research shows that there is four kinds of tides in Persian Gulf including Daily, semiDaily, daily compounded and semidaily compounded tides and on the other regions there is just semidaily compounded ides.The research aims on investigating the effects of climate change parameters (temperature, pressure, and wind rate) on the sea level fluctuations in an annual, seasonal and monthly changes and 20year period in northern coasts of Oman Sea and also its physical justifications. Matarials and methods Research domain from geographical networking point of view, Jask port has longitude and latitude of 570 46' in east and 250 40' in north respectively to Gowatre Bay and in the terminal point of southeast of Iran and southwest of Pakistan has longitude and latitude of 570 46' in east and 250 10' in north respectively. Through calculating mean sea level from tide gauges datum belonged to IOC (Intergovernmental Oceanographic Commission) in the stations of hydrography of Jask and Chabhar ports during 1997 to 2016. Tide gauges for mentioned stations during 24 hours presented 1440 datum. In fact they registered sea level in every minutes. Data presented in every minute through averaging changed firstly into hourly data then into daily and finally into monthly data. In order to compute data based on monthly averages, the tidal effect should be deleted and computed into level fragment that means sea level minus tidal effect. According to the presented information in meteorological organization since 1997 to 2016, data about pressure, temperature and wind force were used in research stations as monthly means and the diagrams used monthly, seasonal and yearly means. In this research, Meta heuresticAlgorithm (Decision Tree Algorithm) The CARD regression tree decomposition algorithm (Classification and regression tree) is used as a type of regression decision tree for prediction purposes. Different elements have been used in simulation using decision tree model. In order to verify the relationship between the final decision tree tree based on the statistical index, graphical graphs and correlation coefficients obtained from the field operation method, visual inspection, ground monitoring and verification of control points were made. Result and discussion . The decision tree model in the Jask area has three parameters: wind pressure and wind speed, and the tree has acted on the basis of these two parameters that the model did not use the temperature parameter in the decision tree and was not selected as an effective parameter Is. Also in Chabahar region, all three parameters are used in the model The abovementioned model has a very high performance in predicting values. In most of the 12month intervals, the model performed its predictions close to real values; in other words, the tree created using data has a good prediction process and can simulate the changes well Also, according to the above figure, the predictions were evaluated. The results indicate that this model can be predicted with high accuracy and 95% confidence level in the region data. Because the temperature parameter has not been able to predict the response variable in the decision tree, the model has been eliminated, and the final equation of Jask and Chabahar is as follows. MSL(Jask) =13.197+5.619 T(1.102)11.092 P (0.195)+7.208 W(0.71) MSL(Chabahar)= 4.520+1.529 T (1.089)1.596 P(0.87)+2.776 W(0.316) Conclusion MSL fluctuations are among the general methods of analysis; therefore, accurate prediction can provide conditions for assessing its status. The purpose of this study was to investigate the effect of data preprocessing on the performance of nonlinear decision tree model in predicting MSL in Jask and Chabahar.The results of this study in all simulations show that pressure and wind parameters are more effective in the final model, which indicates the importance of these parameters in predicting future MSL. Physically, the close relationship between wind speed and water level changes is evident with the strong positive correlation coefficient of the Jask station compared to the Chabahar station in the annual windfall of both regions. Based on the relationship between the final model derived from the decision tree tree algorithm in MSL prediction using available data, it is 95% reliable. Investigating the related geomorphologic forms in the study area, the tidal range fluctuations in the Chabahar region are 1 to 1.5 meters, and in the Jask area more than 3 meters above the time interval are shown well in the diagrams. Therefore, in a closer examination of the processes governing the environments around the coastline, studying and monitoring the status of the tidal region and the influential climatic parameters is necessary.