جستجوی مقالات مرتبط با کلیدواژه "پیش بینی بارش" در نشریات گروه "جغرافیا"

The atmosphere is a fluid and interconnected system. Any change by known and unknown factors in a limited part of the atmosphere can cause larger changes in the general circulation of the atmosphere with the passage of time, this itself causes a change in the water share of each region of the water cycle.The researches that have been carried out in order to predict rainfall in different areas of Iran have mostly focused on atmospheric indicators, which were initially made to predict rainfall or study climate changes in areas other than Iran. In fact, in these researches, it shows the effect of special atmospheric circulation patterns, such as Walker, North Atlantic Oscillation, on rainfall of Iran. On the other hand, pressure changes in limited areas have been analyzed to find a suitable predictor (index) for seasonal rainfall of Iran. It is necessary to carry out complete and comprehensive research (sea level pressure patterns on a planetary scale) to investigate the relationship between sea level pressure changes and precipitation in different climatic regions of Iran.

The study area is considered for research and investigation of the explanation of the changes in the time series of precipitation in the annual time scales by the spatial changes of the sea level pressure in the time lags of one to six months, in the northeastern region of Iran (Khorasan Razavi, Northern Khorasan, and Southern Khorasan provinces).Precipitation data for the 35-year statistical period (1987-2021) on a daily basis for the selected synoptic stations including of Mashhad, Sarkhs, Qouchan, Torbat Haidarieh, Sabzevar, Kashmer, Gonabad, Birjand, Nehbandan, Tabas, Ferdous, Qain, Beshravieh and Bojnoord was obtained from the Iranian Meteorological Organization.In order to investigate the relationship between changes in sea level pressure and annual rainfall in northeastern Iran, monthly data of sea level pressure were used. These data are presented and updated on the website of the National Center for Atmospheric and Oceanographic Studies affiliated with the United States of America since 1979. In general, regression and correlation methods are used to determine the presence or absence of a relationship between variables. The correlation index shows the relationship between two variables that are both affected by common factors. Both of these variables are random and other unknown factors jointly affect their occurrence. Monthly time series of sea level pressure data for the time period from 1987 to 2021 were made for 10512 cells (2.5 by 2.5 degrees of arc). The months of September, August, July, June, May and April respectively correspond to the delays are one to six months. In order to be able to investigate the quantitative and spatial changes of precipitation with sea level pressure in the six months before the onset of precipitation, correlation maps were drawn for each of the six months (time delay of one to six months). In the second step, the significance of the correlation was investigated using inferential statistics (hypothesis testing) and the points whose correlation was significant on the map was drawn In the third step, for the practical use of sea level pressure changes to predict precipitation, the points whose correlation was significant (places with maximum correlation) were selected as predictors (independent variable). Using the data of these selected points and multivariable regression, rainfall forecasting models were fitted. The performance of these models in rainfall forecasting was evaluated using indices such as correlation coefficient, explanatory coefficient and root mean square errors. Models with the best performance in predicting rainfall with a delay of one to six months were selected. The places where the prediction models had the best performance in different months (September, August, July, June, May and April) using the data of those regions were introduced as key regions in the prediction.

The results obtained from the correlation study show that there is a significant correlation between the patterns of changes in sea level pressure and precipitation (12 months from October to September of the following year) with a delay of one to six months. In fitting precipitation models based on sea level pressure changes using selected points from the centers with maximum correlation, the obtained results show that the pressure changes in the eastern and western Pacific Ocean, especially in theeastern ocean, in all annual rainfall forecasting models as a predictor has been selected. Changes in sea level pressure in the Pacific Ocean seem to play an important role in predicting annual precipitation.According to the independent variables, Southeast Asia and regions from the Southeast China Sea and Guam Island to northern Australia can be other important indicator regions in predicting one-year rainfall. In examining the performance of forecasting models in delays of one to six months, it was found that fitting models with a delay of 5 and 6 months have better performance than other models. So that the 5 and 6 month delay models explain 70% and 65% of the rainfall changes, respectively.

The findings of this research show that it is possible to obtain favorable results from changes in sea level pressure patterns in the centers with maximum correlation for rainfall forecasting (12 months from October to September of the following year). Since sea level pressure data are available from various sources, it makes us somewhat independent from the data of physical models of atmospheric circulation. It also provides the possibility that we can have a proper prediction of the amount of precipitation for a year a few months before the beginning of rainfall. Since there is no 12-month rainfall forecast for one-year water resource management policies in the agriculture, industry, drinking water, natural resources and environment sectors, and on the other hand, due to the large range of annual rainfall changes in the northeast Using average rainfall cannot be useful and effective, the results of this research can be helpful.

Assessing and predicting future climate change is of particular importance due to its adverse effects on water resources and the natural environment, as well as its environmental, economic and social effects. Meanwhile, rainfall is also an important climatic element that causes a lot of damage in excess conditions. West Azerbaijan Province is no exception. The aim of this study is to model and predict 30 years of rainfall in West Azerbaijan province. The statistical period studied is 32 years (2019-1987). Selected stations in the province include Urmia, Piranshahr, Takab, Khoy, Sardasht, Mahabad and Mako stations. Average slider time series models, Sarima (seasonal Arima), Health Winters were used for analysis and prediction and also linear regression and Mann-Kendall test were used to determine the data trend. The results show an increasing trend of precipitation in Urmia, Piranshahr, Khoy, Sardasht and Mako stations and a decreasing trend in Takab and Mahabad stations. According to the results of comparing the models used, the Health Winters model with the least error in the absolute mean of deviations, mean squared deviations and the percentage of absolute mean errors was introduced as the best precipitation forecasting model for West Azerbaijan province. province. 

In this study, the results of 24 and 48 hour predictions of mid-scale WRF model with nested slopes with 18 and 6 km separations (implemented in Lorestan meteorology) and with 27 and 9 km separations (implementation) At the Institute of Meteorology and Atmospheric Sciences), without schematic, for a period of 2 months from March 1, 2019 to the end of April 2019 and compared with precipitation observation data for 10 synoptic meteorological stations in Lorestan. For this purpose, 2 * 2 agreement table was used for verification. The results obtained from PC skill score showed that in a period of 24 hours, the ranges of 27, 18 and 9 km in more than 80% of cases were able to occur or not rain. At the provincial level, correctly predict that this score was a minimum of 6% for a range of 6 km. Also, studies performed for a period of 48 hours showed that all slopes in more than 77% of cases showed the accuracy of the occurrence or absence of precipitation. The results of validation in this study have been for days with precipitation and no specific threshold has been considered for precipitation values. It is suggested that in order to improve the relative weakness of the model, validation quantities for specific thresholds (light precipitation, Moderate rainfall and heavy rainfall). The results obtained from the PC skill score showed that in a period of 24 hours, the slopes of 27, 18 and 9 km in more than 80% of cases were able to accurately predict the occurrence or non-occurrence of rainfall in the province. The range of 6 km was minimal and amounted to 67%. Also, studies performed for a period of 48 hours showed that all slopes in more than 77% of cases showed the accuracy of the occurrence or absence of precipitation.* The average quantity of B slope for 4 slopes showed that the number of precipitation forecasts for a period of 24 hours in the surveyed slopes is between 1.39 to 1.49 percent higher than the cases in which precipitation occurred, which indicates that the previous precipitation was higher. The occurrence of precipitation is relative to the occurrence of precipitation. For prediction over a 48-hour period, this quantity is slightly improved and has less error.* The average quantity of TS in the forecast for a period of 24 hours is more than 66% for the ranges of 27, 18 and 9 km, which with the increase of the forecast period to 48 hours, this quantity has improved and in all 4 ranges to more than 72%. it is arrived.* The quantity of FAR in a 24-hour period for all 4 ranges varies from 29 to 38% on average, indicating that only 29 to 38% of the precipitation predictions have not been met. In a period of 48 hours, this quantity has also improved a bit and has reached 21 to 26. The results of quantifying the H collision rate over a 24-hour period indicate that the three slopes of 27, 18 and 9 km had a high ability to predict the occurrence of positive precipitation. Also, in a period of 48 hours, almost all 4 domains had high power.* Examination of the average quantity of F shows that in a 24-hour period for slopes 9, 18 and 27 between 29 to 33% of the cases where no precipitation has occurred, the model has erroneously predicted that this error is 47% for a range of 6 km. Is. In a period of 48 hours, the rate of this error for 4 domains has reached 35 to 50%, which indicates an increase in this type of error with increasing time interval.* Quantity of Pierce PSS skill score in 24-hour period shows 39 to 69% improvement for the studied slopes, among which the 9 and 27 km slopes have better performance. In a period of 48 hours, the value of this quantity has reached 48 to 61%, which, unlike the period of 24 hours, is the best value in the range of 6 km. In order to evaluate the model more accurately, further case studies are suggested in different seasons of the year.* Increase model execution time.* The model should be executed with different schemas.* The results of validation in this study have been for days with precipitation and no specific threshold has been considered for precipitation values. It is suggested that in order to improve the relative weakness of the model, validation quantities for specific thresholds (light precipitation) , Moderate rainfall and heavy rainfall).* The results of this research should be compared with GFS and ECMWF models.

In today's world, rapid and sustainable development is the ultimate goal of all countries. The main constraint to sustainable development is the climatic conditions of countries, including rainfall, which is most correlated with natural disasters. One of the methods used for prediction is deep neural networks. The main purpose of this study is to increase the accuracy of hourly rainfall forecasting in Khorasan Razavi region using one of the deep neural network methods. In this study, we present a deep neural network architecture using the Ron Neuron-Based Random Neural Decomposition Stochastic (RSDSAE) method for short-term precipitation prediction. In order to improve the accuracy, rough neural networks (RNNs) contribute to the rainfall uncertainty and to improve the speed and accuracy, the combination of the sparse algorithm with the above model has been used and We also used the predicted output data output of the WRF model, and tests are calculated on rain data and by two criteria of RMSE, MAE and with five self-esteem, respectively, 0.7912 and 0.7662, and has been able to better performance than the model (RSDAE) of itself Show. Keywords: Deep Neural Network, Precipitation Prediction, stacked denoising sparse autoencoder.

Rainfall is a climatic factor that affects many human activities such as agriculture, construction, electricity generation, etc. Therefore, having a proper method for predicting rainfall can take preventive and mitigating measures for natural disasters. Over the past few years, deep neural networks (DNNs) have been used as a successful mechanism for solving complex problems in areas such as machine vision, image recognition, etc. (Hern´andez, et al., 2016). Deep neural network is a set of multi-layered architectures that have been trained using unsupervised algorithms and have challenges in training deep neural networks such as slow network training, overprocessing, and so on. Most existing neural network models have three drawbacks: (Khodayar, et al., 2017)1- Most of the architecture is shallow.2. Some methods require handmade engineering features that are tedious.3- Most methods do not have direct knowledge about rainfall uncertainty.In this study, to solve problems one and two, we combined the proposed method of stacked denoising sparse autoencoder (SDSAE) and to solve problem three, we combined the proposed method with Neuron Rough.2.

2-1. Data and area of study:In this study, the study area is Khorasan Razavi province and the output of networked data is the WRF model, which is a mid-scale regional forecasting model. We considered eight days of regional precipitation hourly data as target variable data for training and testing algorithms. We prepared the data used from the output of two regional forecast model implementations (mid-scale atmospheric forecast model) on 26 and 28 October 2018 for each implementation with 4 days forecast.2-2. Steps to perform the proposed method:The proposed method has five main steps:1) Receive WRF output prediction data,2) Apply noise to input data.3) Validation of deep network parameters.4) Unsupervised learning with RSDAE to build deep multilayer network.5) Learning by monitoring (fine tuning) by BP algorithm with SGD method2-3. Evaluation criteriaTo evaluate the proposed method SRDSAE and RSDAE, we use two criteria, mean squared error (RMSE), absolute mean error (MAE).

Predicting rainfall is one of the most important and fundamental challenges for researchers. Properly predicting rainfall improves the lives of the general public and even the proper planning of governments for the correct use of rainfall. Various algorithms and methods have been proposed for forecasting issues in recent years. Deep learning methods are one of the most popular methods for predictive problems. In this research, we have used the stacked denoising autoencoder method to predict rainfall and combine this method with neuron-rough and sparse algorithm to increase the accuracy of the forecast. In this research, we applied the proposed method on the output data of the WRF mid-scale forecast model in Khorasan Razavi province. The proposed RSDSAE(Rough Neuron Based stacked denoising sparse autoencoder) method and RSDAE (Rough Neuron Based stacked denoising autoencoder) method have been compared and analyzed for better comparison in precipitation forecasting based on two evaluation parameters RMSE and MAE. The proposed method has been able to reduce the forecast error by reducing the RSDAE network and subsequently reduce the RMSE and MAE criteria, or in other words, improve the rainfall forecast.

Precipitation plays a decisive role in meeting the water needs of various crops, dam reserves, feeding surface and groundwater resources, and the occurrence of floods and droughts. Lack of access to long-term daily rainfall data and the high cost of setting up ground meteorological stations necessitate the replacement of low-cost methods with high-precision, high-volume data such as satellite data. The aim of this study was to evaluate the accuracy of precipitation data predicted by TRMM and GPM satellites in Bandar Abbas metropolis during a 17-year period from 2000 to 2017. Hourly precipitation data of TRMM and GPM satellites were obtained from databases. After analyzing the data format in MATLAB environment, hourly precipitation information was extracted. The results showed that the accuracy of both TRMM and GPM models in precipitation forecasting was appropriate and close to each other; it was often overestimated so that 75% of TRMM model precipitation forecasts and all forecasts GPM models were overestimated. The results showed that the TRMM model was more accurate than the GPM model in accurately predicting the occurrence of rainfall events and had less error in predicting unrealistic rainfall and the highest accuracy of the TRMM and GPM models is on a monthly, annual and daily scale, respectively. The value of EF index in TRMM model varies from -284.52 to 0.71 and in GPM model from -25514 to -1.25. The value of the EF index in the TRMM model predictions was positive in 42% of events, while, in the GPM model, it was not positively predicted in any event. The general conclusion of the research is that TRMM satellite is a suitable tool for monitoring and forecasting precipitation.

 

Rain is one of the most important atmospheric phenomena affecting human life. Rainfall prediction is important for various purposes such as planning for agricultural activities, forecasting floods, monitoring drought and providing resources for consumable water. The rapid expansion of using artificial neural networks (ANNs) as an experimental and efficient model in various sciences including meteorology and climatology implies the high value of studying these types of models.

The purpose of this paper is to predict the monthly rainfall in Iran, using the combination of artificial neural networks and extendedKalman filter. For this purpose, the monthly average rainfall data of about 180 synoptic stations spreading across the country were used during the years 1951 to 2016, then, the monthly rainfall was predicted for the year 2017 using the article’s method.  Artificial neural networks are a method for the approximation of the functions and prediction of the future state of various systems. Artificial neural networks discover the law latent in them and transfer it into the network by processing the experimental data. The smallest processing unit of information in the artificial neural network is neuron that builds the bases for the application of neural networks. Each neural network consists of a number of nodes which are the neurons, and the communication weights that connect the nodes together. Input data is multiplied by their corresponding weights, and their sum is entered into the neurons. Each neuron has a transfer function. This input data passes through the transfer function and specifies the output value of the neuron. The back propagation algorithm is one of the most commonly used algorithms for teaching these networks, but the back propagation algorithm depends on the selection of the number of hidden neurons. Also, the convergence speed of the back propagation algorithm is very slow comparing with the proposed method in this paper, and is very sensitive to the noises present in the input and output data set, which is used for teaching the neural network. In addition, it has a poor performance in modeling the complex processes. One of the most famous methods to eliminate the aforementioned defects is the use of the Kalman filter. The Kalman filter contains a set of mathematical equations that performs a repeated process, prediction and updates, and is also an optimal tool in minimizing the covarianceof the estimated error. The leading neural network can be considered as a nonlinear dynamic system with synaptic weights and equate the teaching of the neural network with the problem of estimating the state of the nonlinear system. Therefore, the extended version of the Kalman filter can be used to estimate the adjustable parameters of the artificial neural network like the neural network weights.

The climatic zonation of Iran was used for the data separation by the method of Coupon-Geiger which has been conducted by other researchers, and Iran was divided into eight climatic zones. This zonation divides Iran into dry and hot desert, dry and cold desert, dry and hot semi-desert, dry and cold semi-desert, moderate with dry and hot summers, rainy moderate with warm summers, snowy with dry and hotarm summers, snowy with dry and warm summers climates. This artificial neural network which has been taught by the extended Kalmanfilter, was used for the prediction in each of the climatic zones. A multi-layered artificial neural network with two hidden layers has been used. There are 10 neurons in each of the hidden layers, and 7 neurons in the input layer, which are the same numbers as the network inputs. The methodology is that the year and number of months, the average monthly temperature, the average monthly wind speed and the geographic location of the synoptic stations are considered as the input parameters, and the average monthly precipitation as the output parameter. The difference between the observed and the predicted values of the monthly precipitation in 2017 was calculated ​​for all stations and was considered as an error. The Root Mean Square Error (RMSE) of these differences was calculated for the 8 climatic zones. The RMSE is lower for dry and hot desert climate than for dry and cold desert climate. This RMSE is lower for dry and cold semi-desert climate than for dry and hot semi-desert climate. The RMSE is lower for moderate climate with dry and hot summers than for moderate rainy climate with warm summers. The RMSE is lower for snowy climate with dry and hotsummers than for snowy climates with dry and warm summers.

In most cases, the RMSE for hot climates is less than others that represents the efficiency of the proposed method for the prediction of  monthly precipitation in hot climates.

There are several ways to study drought. Method of analysis rainfall data, Public Sector analysis methods is drought. Therefore, accurate prediction and before the outbreak precipitation could provide the conditions for assessing the drought situation. The purpose of this study is investigating the effect of data preprocessing on the performance of the decision tree model to predict drought in synoptic station in Sanandaj. In this study, CART algorithms (Classification and regression tree) has been used as variety of decision tree regression in order to predict precipitation forecast of12months. The data used in this study are the monthly precipitation, relative humidity, the maximum temperature, the average temperature, wind direction and wind speed in a specific statistical period(1970 - 2010). To assess the created trees in this study, different statistical measures have been used which in the end results show that in synoptic station in Sanandaj, decision tree regression model is a relatively efficient model to predict drought in which using a moving averages compared to other states led to Increasing the efficiency of decision tree mode land providing thread just mint in the range of changes, the input data with a high reliability is able to estimate the amount ofprecipitation12months before it occurs which in the simulation carried outing this study, when the five-year moving average of the data has been used to implement the model ,combination of previous rainfall, maximum temperature has been identified as the most appropriate states. The findings shows that applying moving average to the original data, dramatically improves the performance of the model. In these circumstances, the decision tree method regression in Sanandaj station with high reliability level estimate the occurrence of precipitation in 12 months ago.

Precipitation is one of important parameters of climatology and atmospheric science that have more importance in human life. recently، extensive flood and drought entered many damage to most parts of the world. Precipitation forecasting and alerts management role is responsible for these problems. Today، Artificial Neural Networks are one of developed method that applied for estimate and predict of parameters that using Intrinsic relation among data. The purpose of this study is prediction of daily precipitation using daily data of common statistical period، 23 years، (2012-1989) from meteorological stations، by Perceptron Neural Networks and Radial Basis Function in Kerman، Baft، Miandeh Jiroft. Different combinations of parameters، such as minimum temperature، maximum temperature، average temperature، relative humidity، wind speed and direction and the mean pressure as input for Artificial Neural Network and precipitation was considered as the output of the Network. The analysis of output results from، Radial Basis Function shown that these models have better accuracy and a high ability to forecast precipitation than Perceptron Neural Networks. The analysis of output results from، Radial Basis Function Neural Networks shown that these models have better accuracy and a high ability to forecast precipitation than Perceptron Neural Networks. such that in the best combination with minimum and maximum parameters، minimum temperature and relative humidity، the wind speed and direction and mean of pressure in Kerman station with correlation coefficient 0. 907 and root of mean square error of 0. 014 known as the best model for predicting rainfall in this paper.

In this study، it has been tried to test the accuracy of output and hence the model efficiency in the predictions of 24 and 48 hour precipitation forecasting by choosing different schemes be assessed. Therefore، in order to test the accuracy of model product، two conditions of rainfall precipitation and lack of rainfall precipitation were considered. Four thresholds were taken. Adaptive table and threat scores were calculated. The final results of the analysis are summarized. The results indicate that among different configurations، the model output for the two configurations of GDMYJ and KFMYJ which in both of them boundary layer scheme Mellor-Yamada-Janjic were used، are closer to the truth and their accuracy are more acceptable. Then proportion correct (PC) for the four thresholds، without precipitation (≤ 1. 0)، less than or equal to 1 mm (1 ≥ and> 1. 0)، less than or equal to 10 mm (10 ≥ and> 1) and more than one، and more than 10 mm (<10) for 48 hours predictions have acceptable values such that totally about 80 percent، predicting the occurrence or non-occurrence of precipitation is done properly. Amount of H represents for two thresholds means less than 1. 0 and more than 10 mm is close to one، it means that the model for the two mentioned thresholds has higher accuracy. This quantity for the range of one tenth to a millimeter is 0. 3 which indicates the relative weakness of model in this range

The aim of this research is investigating the relations between climatic large scale signals and spring rainfall of Khorasan-e Razavi province. In this research, we have analyzed 38 years of rainfall data in khorasan-e Razavi province, located in northeastern Iran. We attempted to train Artificial Neural Network based upon climatic large scale signals with 38 years of rainfall data. For performance evaluation, network predicted outputs were compared with the actual rainfall data. At the outset of this study, the relationships between synoptically pattern variations, including Sea Level Pressure (SLP), Sea Surface Temperature (SST), Sea Level Pressure Gradient (?SLP), Sea Surface Temperature Difference (?SST), Air Temperature at 700 hPa, Thickness between 500 and 1000 hPa level, Relative Humidity at 300 hPa and Precipitable water are investigated. In the second step, model was calibrated from 1970 to 1997. Finally, rainfall prediction is performed from 1998 to 2007. Simulation results reveal that Artificial Neural Network is promising and efficient. Root mean square was obtained 2.5 millimeters.

Since immemorial past, drought has been one of the dangerous natural disasters for human life. As it is seen in historical texts, this phenomenon has always caused many social-economic changes, such as wars, famine and immigration. In this research, the wet and dry periods of some of meteorological stations of Lake Urmia Basin, located in the northwest of Iran, have been analyzed. For this purpose, first the stations of Urmia, Tabriz, Saqez and Miandoab have been normalized using yearly methods. In other words, the precipitation was turned into a standard score (Z-Score). Then, the intensity of either of dry and wet periods was defined in two classes: weak to average dry (-1.5 <z <-0.5), severe dry (z<-1.5), weak to average wet (+0.5 <z <+1.5) and extreme wet (z> +1.5). In addition, the roughly normal period (-0.5 <z <+ 0.5) was considered as a period neither significantly wet nor suffering from drought. In the next step, a Run Test was performed on monthly precipitation data and their homogeneity was confirmed. Subsequently, using time series method (SARIMA), the monthly precipitation of these stations was modeled and predicted for the future (based on the most appropriate model) until 2002. Finally, according to the predicted values, the degree of severity of the year’s wet-or-dryness has been determined. The results of the study show that the annual precipitation of this basin in the years 2001 and 2002 is almost normal and will be accompanied by an upward trend and exit the state of drought.