• <tr id="yyy80"></tr>
  • <sup id="yyy80"></sup>
  • <tfoot id="yyy80"><noscript id="yyy80"></noscript></tfoot>
  • 99热精品在线国产_美女午夜性视频免费_国产精品国产高清国产av_av欧美777_自拍偷自拍亚洲精品老妇_亚洲熟女精品中文字幕_www日本黄色视频网_国产精品野战在线观看 ?

    CDLSTM:A Novel Model for Climate Change Forecasting

    2022-08-24 03:26:40MohdAnulHaq
    Computers Materials&Continua 2022年5期

    Mohd Anul Haq

    Department of Computer Science,College of Computer and Information Sciences,Majmaah University,AL Majmaah,11952,Saudi Arabia

    Abstract: Water received in rainfall is a crucial natural resource for agriculture, the hydrological cycle, and municipal purposes.The changing rainfall pattern is an essential aspect of assessing the impact of climate change on water resources planning and management.Climate change affected the entire world, specifically India’s fragile Himalayan mountain region, which has high significance due to being a climatic indicator.The water coming from Himalayan rivers is essential for 1.4 billion people living downstream.Earlier studies either modeled temperature or rainfall for the Himalayan area;however, the combined influence of both in a long-term analysis was not performed utilizing Deep Learning(DL).The present investigation attempted to analyze the time series and correlation of temperature (1796–2013) and rainfall changes(1901–2015)over the Himalayan states in India.The Climate Deep Long Short-Term Memory (CDLSTM) model was developed and optimized to forecast all Himalayan states’temperature and rainfall values.Facebook’s Prophet (FB-Prophet) model was implemented to forecast and assess the performance of the developed CDLSTM model.The performance of both models was assessed based on various performance metrics and shown significantly higher accuracies and low error rates.

    Keywords: Himalayas;climate change;cdlstm;facebook;temperature;rainfall

    1 Introduction

    Climate change affected India,specifically the fragile Himalayan mountainous region,which has high significance due to climatic indicators,the origin of rivers,municipal,agricultural,hydroelectric purposes, minerals, and tourism [1–3].The Himalayan area is the world’s highest mountain system that stretches over 2,500 km,covering twelve states of India either partially or fully.These are Jammu and Kashmir (J&K, now Jammu, Kashmir, and Ladakh), Himachal Pradesh (Him), Uttarakhand(UK),Nagaland,Manipur,Mizoram,Tripura(NMMT),Assam,and Meghalaya(A&M),Arunachal Pradesh (Arun) and sub–Himalayan West Bengal and Sikkim (WB&S).The high-altitude area of the Himalayas is covered with snow and glaciers.The water from snow and glaciers is essential for municipal,agricultural,and hydroelectric[1–3].The Himalayan area impacts the climate of the entire Indian subcontinent by controlling the monsoon and rainfall patterns.The Himalayan ecosystem is very fragile and sensitive to climate change parameters such as temperature and rainfall.The criticality of rainfall analysis is evident from the June 2013 Uttarakhand disaster.The UK state received 847%extra rainfall in the third week of June,resulting in the Kedarnath flood disaster that took more than 5000 lives.

    In current times, the utility of machine learning (ML) approaches has a significant presence in almost every area, and the successful utilization of deep learning (DL) has opened new dimensions for efficient time series forecasting.ML/DL has been used extensively for climate analysis and forecasting[4–9].A study by[10]provides a comprehensive report on climate change in the Himalayas and suggests modeling future climate status,especially for the Himalayan region.[11,12]applied Dl techniques such as DNN and RNN for weather forecasting; however, the scope and data used in their investigation were limited.However,significant issues with these studies are rigorous parameter tuning, cross-validation of the model on different data, and computational efficiency.Additionally,while previous studies have addressed annual climate issues for the entire expanse of India,very few studies have explicitly focused on all the Himalayan states due to the complexity of data for the Himalayan region[13–15].Understanding the climate variability for Himalayan states on a monthly temporal scale is crucial for hydrological and climatic models [16].The present study focuses on all the Himalayan states to provide comprehension of changing temperature and rainfall patterns.The present investigation contributes to three significant aspects:detecting rainfall and temperature trends, analyzing the correlation between temperature and rainfall, and forecasting the temperature and rainfall using a novel CDLSTM model and Facebook Prophet(FB-Prophet)Model.

    1.1 LSTM

    LSTM networks are a type of Recurrent Neural Network(RNN)that uses special units(cells)and standard units to overcome the limitation of traditional RNN [17–20].There are three gates, which are contained by a cell in LSTM.The first gate is the input gate,the second is termed the forget gate,whereas the third is the output gate.The LSTM network composition function’s description is based on the input node,and the three gates are contained by a cell,cell state,and output layer.Eqs.(1)–(7)are as follows[20,21].

    Input node

    Input gate

    Forget gate

    Output gate

    Cell state

    Hidden gate

    Output layer

    1.2 Facebook’s Prophet Model

    Facebook’s Prophet is an open-source forecasting tool based on a decomposable additive model,similar to a generalized additive model(GAM).Prophet can fit nonlinear time series with seasonality.The Prophet forecast model can be expressed as Eq.(8).

    where,F(xiàn)(t)=forecast,L(t)=long term trend,S(t)=short term trend,E(t)=error,and X(t)=any other influencing variable to forecast.

    Prophet has two models:logistic growth model(LGM)and piece-wise linear(PWL)model.The selection of the model depends on the time series data.The LGM model can be used if the time series shows non-linearity,saturation,and no change after reaching the saturation point.If the time series exhibits linear tendency and a previous track of shrink and growth,then PWL is a better option.The LGM can be expressed as Eq.(9).

    where CC=carry capacity,g=growth rate,and o is an offset parameter.The PWL can be expressed as Eq.(10).

    2 Dataset

    2.1 Rainfall

    The monthly rainfall dataset was obtained from more than 3000 rain-gauge stations spread over India, covering 115 years (Jan 1901–Dec 2015).The dataset was released by Indian Meteorological Department(IMD)(https://www.imdpune.gov.in/).

    2.2 Temperature

    The Berkeley Earth monthly average data from Jan 1796–Aug 2013 was procured from https://gui des.lib.berkeley.edu/publichealth/healthstatistics/rawdata.It was generated based on a variety of data,including bias-corrected station data,regional data.The data was developed from various sources with quality control,and monthly averages were created from daily data.A standard temporal observation period of Jan 1901–Aug 2013 was considered to understand the relationship between temperature and rainfall.

    3 Methodology

    3.1 Data Preprocessing

    For in-depth analysis of seasonal patterns of temperature and rainfall, the data was divided into four seasons based on India’s meteorological and international standards, i.e., Dec-Feb as winter, March to May as spring, June-Sep as monsoon, and Oct-Nov as post-monsoon or autumn.

    Temperature and rainfall data were used;therefore,the term monsoon was used instead of summer for rainfall analysis purposes.Data transformation is crucial before implementing any ML model.Three data transformations were applied in the current investigation.The first transformation was removing missing values and replacing them with average values from the respective records.The second step was transforming time-series data into input and output so that the output of a step could become the input for the next step to forecast the value of the current time step.As described earlier,the total common data in the time series covered 1352 monthly values.The first 980 months’dataset for all Himalayan states was taken for the training, while testing took 240 months, and validation used the dataset of 120 months of the LSTM model; the remaining twelve months of data were kept separate from the training process for the unbiased external validation of the LSTM prediction.The third transformation was the scaling of time series data from–1 to 1.These three transformations were inverted after the prediction step to get the values at the original scale so that the uncertainty calculation could be adequately assessed.

    3.2 Trend Analysis

    Mann-Kendall tests [22,23] were carried out for trend analysis, detecting trends and changes in temperature and rainfall over the years of analysis.Sen’s slope values[24]were used to understand the trend of GWSC change for all Himalayan states from Jan 1901 to Dec 2015.Statistics of Mann-Kendall S value[22,23]were evaluated for chronologically placed observations in the time series Eq.(11).The observations VAR(S)variance in the time series was also estimated as per Eq.(13).Standardized test Z Eq.(14)[25]for the statistical analysis was also performed.

    Here,XiandXjare chronologically placed values of variables in the time series,n represents the total count of observations, ties for pth value is shown astp, and tied values number is shown as q.When Z is positive,it means an increasing trend in the climatic variable and vice versa.

    3.3 Correlation Analysis Between Variables

    An attempt was made to study the correlation analysis based on Moment Correlation Coefficient(MCC)among temperature and rainfall values for all the twelve Himalayan states from Jan 1901 to Aug 2013,as per the availability of a common temporal dataset.

    The MCC summarizes the direction and degree of linear relations between actual and modeled datasets.The correlation coefficient can take values between–1(perfectly negative correlation)through 0(no correlation)to+1(perfectly positive correlation).The MCC formula to compute the correlation coefficient is given in Eq.(15).

    Here,N represents the number of pairs of data.The terms X and Y are parameters.

    3.4 Development and Tuning of CDLSTM Model

    Keras library with TensorFlow and Python version was used to develop the LSTM models in the current study.The libraries used in the current investigation were Plotly, NumPy, Seaborn, Pandas,Matplotlib,and scikit-learn.A four-step procedure was applied to develop the LSTM model.

    The first step was to define the LSTM network to aid LSTM model development.Eight LSTM layers were used in the current investigation, in which four layers were dense, and three were the dropout layers,see Fig.1.The dropout layer“drops out”inputs to a layer,which may be input variables from a previous layer.A value of 0.5 was chosen with two dropout layers.The second step was the network compilation.It required several parameters,such as an optimization algorithm to train the network and the loss function to evaluate the network.Several optimizers were tested based on their performances.The third step was the fitting of the LSTM model.The fourth and essential step was the prediction using the LSTM model.We forecasted the output step by step for the test data.The model fed the current forecasted value back into the input window by moving it one step forward to forecast the next step using the moving-forward window technique [26].Here we used a moving forward window of size 12.We forecasted the average temperature individually for all Himalayan states from Sep 2012 to Aug 2013 using one step ahead regression based on window size.

    Several hyper-parameters such as optimizer, number of units, learning rate, momentum, and activation functions must be chosena prioriand then tuned based on the RMSE values.Tuning the hyper-parameters of any neural network model is essential for evaluating the performance and stability of the DL model.The first configuration tuned was the number of nodes,which affected the LSTM model’s learning capability.A higher number of nodes ensures excellent learning ability for complex data at the cost of computation time and can cause overfitting.Different nodes(2,4,6,8,and 10)were tested for different configurations.A lesser average RMSE value of 1.4 and the lowest variance based on 20 experimental runs were obtained with four nodes.However,since it could indicate overfitting,dropout was applied to prevent overfitting, where the neurons were randomly chosen and ignored during model training to address the issue of overfitting.The number of epochs(10,20,30,40,and 50),optimization algorithm(Adam,RMSProp,Adagrad,SGD),and individual learning rate(1e-2 to 1e-6)were also rigorously tuned.It was observed that the tuned LSTM model with eight nodes,trained for 20 epochs with an ADAM optimizer having LR of 1e-2,showed the best performance based on RMSE and computational efficiency in the current investigation.

    Figure 1:The architecture of the CDLSTM model used in the current investigation

    3.5 Facebook’s Prophet Model

    The Prophet forecast model looks straightforward;however,the computation can be complex due to the selection of parameters.The selection of the LGM or PWL model depends on the time series data.The LGM model was applied for rainfall data due to the time series;however,PWL model was applied for temperature forecasting as it exhibits linear tendency.

    The uncertainty in the forecasting values can be obtained by forwarding the GAM model,which can be expressed as Eq.(16).

    where?is a rate scale parameter.

    The FB-Prophet model was imported.The Prophet model was fitted with training data, and forecasting was implemented based on 12 periods and month start(MS)as frequency.

    3.6 Uncertainty Assessment of the LSTM Model

    The MCC, Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Nash-Sutcliffe coefficient(NSE)were utilized to evaluate the uncertainty of the LSTM model output.The Mean Absolute Deviation (MAD) was also considered to analyze the LSTM model’s accuracy between measured and predicted values.

    RMSE is a method to calculate the error or accuracy in predicting models based on standard deviation Eq.(17).The final output is given in the form of the standard deviation of the error’s magnitude,as per Eq.(17);the individual calculations are outputted as residuals based on[27].

    Here,P_i is the ithLSTM predicted value,andOiis the ithoriginal value.

    The MAPE method was used to calculate the prediction accuracy of the LSTM forecast.The calculation was based on the difference between the original values and values forecasted by the LSTM and dividing the original value difference.It was then multiplied by the number of observations and 100 to obtain the percentage error(18)[28].

    Here,Atrepresents actual value.Similarly, symbolFtrepresents the forecasted value or the predicted value.MAD was used to calculate the dispersion of LSTM forecasted values,as per Eq.(19).A lower value of MAD indicates that the forecasted data values are closely concentrated.

    where Piis the ith data value,?is the mean value,and n is the number of samples.

    The NSE or efficiency coefficient test determines the magnitude between the residual time series and variance of actual data, and its value ranges from –∞to 1, see Eq.(20).An output near one indicates higher model quality and reliability,while a value below zero suggests unreliable model.NSE test has been utilized in LSTM and FB-Prophet forecasting models[29].

    wherey,andyfare the actual time series, mean of the actual time series, and forecasting series,respectively.

    4 Results and Discussions

    4.1 Trend Analysis of Temperature and Rainfall

    The monthly precipitation was found to have decreased over the period 1901-2015 in the Himalayan states.A 20 mm decrease was observed from 180 mm to 160 mm.The decrease in precipitation occurred after July 1995.The highest monthly precipitation (742 mm) was received in July 1948.Feb 2005 has the lowest average temperature for the UK,Him,J&K,A&M.The highest avg temperature was 28.28oC for A&M for June 2013,followed by NMMT(27.87oC)the same month.The highest rainfall(1347.2 mm)occurred for NMMT among all Himalayan states in Aug 1969,followed by A&M(995.2 mm)for July 1984.Higher temperatures were increased after the year 2000,and the occurrences of high rainfall were decreased after the 1990s.It was evident that climate is changing rapidly,especially in Himalayan states.Results with a confidence factor ≥of 90%indicate a significant trend in the rainfall averaged over the Himalayan states,see Tab.1.Fig.2 represents the rainfall time series for all 12 Himalayan states, with subplots shown as (a) J&K; (b) Him; (c) UK; (d) A&M; (e)WB&S;(f)Arun;(g)NMMT.It is worth noting that both Arun and NMMT showed decreasing trends for rainfall.However, rainfall over J&K, UK, WB&S showed an increasing trend.Interestingly, the Himalayan state Arun showed high average rainfall while the lowest for J&K,see Fig.2.

    Table 1:Change in rainfall(1901–2015)

    The mean annual temperature of Himalayan states was observed to have increased around 1.07°C between 1796–2013.Remarkably,it increased only by 0.98°C for the entire of India for the same period.The temperature of the Himalayan states is increasing faster than in the rest of the country.The average winter temperature rose by 1.27°C over the past century,while post-monsoon temperature increased by 1.03°C,see Tab.2.The intense increase in temperature occurred after July 1998.The temperature of monsoon and spring did not show a significant difference for West Bengal from 1796–2013.

    4.2 Correlation Analysis Between Temperature and Rainfall

    The MCC was performed to understand the relationship between temperature for the twelve Himalayan states from Jan 1901 to Aug 2013 as per the common temporal dataset availability Eq.(16).There was a strong correlation (0.98) between the average temperature of all Himalayan states, see Fig.3.Because the average temperature showed an increasing trend in all Himalayan states.It was necessary to understand the influence of temperature on rainfall.The correlation coefficient between temperature and rainfall was significantly strong for Northeastern Himalayan states A&M (0.80),WB&S (0.78), NMMT (0.76), and Arun (0.62); however, it was weak for Northwestern Himalayan states UK(0.5)and Him(0.39)and J&K(0.18).The stronger correlation in the northeastern states is due to an increase in temperature and decrease in rainfall;however,northwestern states such as J&K and UK showed an increase in rainfall and temperature in an inconsistent pattern.The primary reason for the increase in rainfall is the complex assimilation of monsoon and westerlies in the northwestern Himalayan region.

    Figure 2:The average rainfall for different seasons in all Himalayan states from Jan 1901–Aug 2013.X-axis represent years and y-axis represent rainfall in(mm)

    Table 2:Change in Temperature(1901–2015)

    Figure 3:Correlation heatmap of climate variables for all Himalayan states

    4.3 Temperature Forecasting Based on CDLSTM and FB-Prophet

    There was a possibility while predicting the future values that the LSTM and FB-Prophet models’output may be uncertain as the model’s output was fed back into it as input.Therefore,we forecasted the temperature from Sep 2012 to Aug 2013 and compared them with the actual values based on the coefficient of determination, RMSE, MSE, MAPE, MAD, and NSE, see Tabs.3 and 4.The developed LSTM model was likely to estimate the possible future values of temperature accurately,given its reliability, see Figs.4 and 5.The loss of the proposed CDLSTM model for corresponding epochs showed in Fig.4.The temperature was forecasted for all twelve Himalayan states.The loss of the CDLSTM model indicated that the model achieved significantly lower loss after the first epoch.Fig.5 shown temperature forecasting using developed novel CDLSTM model and FBProphet for four Himalayan states; J&K, Him, UK, and NMMT.Both CDLSTM and FB-Prophet model’s performance showed good forecasting values for all months, including Jan 2013, where the temperature was low due to the peak winter season.

    Table 3:Performance evaluation of developed CDLSTM model based on different accuracy metrics

    Table 4:Performance evaluation of FB-Prophet LSTM model based on different accuracy metrics

    Table 4:Continued

    Figure 4:Loss vs.no of epochs for developed CDLSTM model for temperature forecasting

    4.4 Rainfall Forecasting Based on CDLSTM

    The developed CDLSTM model was used to forecast the rainfall values for all Himalayan states.The training and testing performance of the CDLSTM model is shown in Fig.6, it represents the forecasted rainfall for all 12 Himalayan states,with subplots shown as(a)J&K;(b)Him;(c)UK;(d)A&M;(e)WB&S;(f)Arun;(g)NMMT.It is worth noting that the CDLSTM based forecasted rainfall shown good matching with actual rainfall for all Himalayan states except J&K and Him.J&K and Him showed less coherence with forecasted rainfall due to the high complexity of the snowfall and rainfall pattern due to Indian monsoon in the summer season and westerlies in the winter months.The CDLSTM model forecasted the temperature for the entire time series as training and testing data from Jan 1901 to Dec 2015 and compared them with the actual values based on the coefficient of determination,RMSE,MSE,MAPE,MAD,and NSE,see Tab.5.The CDLSTM model for rainfall forecasting showed less accuracy for J&K,which may be due to the inconsistent rainfall pattern for the J&K state.Best forecasting values were obtained for A&M state using the developed CDLSTM model based on all performance metrics, see Tab.5.An interesting observation was that the performance of the developed CDLSTM model was significantly better for temperature forecasting than rainfall forecasting.The primary reason behind this difference was the higher fluctuation in rainfall data than temperature data.

    4.5 Comparison with Other Studies

    The forecasting performance of the current study was compared with other benchmark studies based on R2and RMSE values,see Tab.6.Based on performance metrics,the present study’s models showed better results than previous studies.

    Figure 5:Performance of CDLSTM and FB-Prophet models for temperature forecasting using novel CDLSTM and FB-Prophet forecasted for(a)Jammu and Kashmir,(b)Himachal Pradesh,(c)Uttarakhand,(d)Nagaland Manipur Mizoram and Tripura

    The FB-Prophet model implemented in the present investigation with the PWL algorithm showed remarkably efficient performance based on accuracy metrics, see Tabs.3 and 4.As per available literature,the current investigation’s performance achieved by the FB-Prophet model for temperature and rainfall forecasting is the highest,based on accuracy metrics.The developed CDLSTM model has lower accuracy than the FB-Prophet model;however,the CDLSTM model showed better performance than models applied in previous studies.In the present investigation, the seasons in the Himalayan area were defined as per IMD and international standards.It is an important criterion to put the correct months as per the respective seasons.[4]defined summer as Jan and Feb,which might be an error.For India, the monsoon season overlaps with summer; therefore, studies focusing on rainfall trends and forecasting should use monsoon instead of summer rainfall to reduce ambiguity.Models based on ML require intense hyperparameter tuning to achieve performance with model stability.ML models might provide higher accuracy without proper optimization; however, this accuracy might be illusionary and unstable.The study was done by [4] to show a lack of hyperparameters tuning.The present investigation attempted rigorous hyperparameter optimization to ensure efficient model performance with model stability for the developed CDLSTM.Additionally, the CDLSTM model developed on temperature dataset was applied and assessed on a different dataset,i.e.,rainfall dataset.In order to evaluate one model, it is imperative to conduct a comparative analysis with a different model for secular evaluation.The present investigation compared the developed CDLSTM model with the popular FB-Prophet model and showed significant performance;however,[4]did not compare the ANN model with another model.Data preprocessing steps such as removing missing values, data transformation,etc.,are vital to building an efficient ML forecasting model.The study[4]observed that the rainfall data from 1901-2015 has no missing values;however,the present investigation found that the same dataset for the entire India had 1036 missing values.Therefore,the present investigation replaced the missing data values with the mean value of the respective parameter.

    Figure 6:Performance of CDLSTM model for rainfall forecasting, training, and testing forecasted rainfall,shown using red,green,and blue,respectively

    Table 5:Performance evaluation of the developed CDLSTM model based on different accuracy metrics

    Table 5:Continued

    Table 6:Comparison of temperature and rainfall forecasting/analysis with other studies

    4.6 Computational Efficiency of the Present Investigation

    After optimizing the CDLSTM model,it took 45 s 21 ms/step for 20 epochs,i.e.,a total of 907 seconds or 15 min 12 sec to complete the training.Optimization saves computation cost by selecting the best number of parameters,including the number of epochs.The optimized model took only 40%computational time compared with 50 epochs in 40 minutes.The imported FB-Prophet model took three minutes to perform the results,only 20%of the computational processing time.

    5 Conclusions

    The present investigation provides an understanding of the long-term historical and forecasted data of temperature and rainfall for India’s Himalayan states.A DL-based LSTM model was developed based on rigorous hyper-parameters tuning to forecast the temperature and rainfall.The correlation coefficient, MSE, RMSE MAPE, NSE, and MAD were obtained to evaluate the CDLSTM model performance.All the twelve Himalayan states showed increasing temperatures after 2000 and a decrease in rainfall after 1990.Arun and NMMT showed decreasing trends for rainfall;however,rainfall over J&K, UK, WB&S showed an increasing trend.The Himalayan state with the highest average rainfall was Arun,while the lowest average rainfall was for J&K.Mean annual temperature of the Himalayan states increased around 1.07°C between the last two centuries; interestingly, it has increased 0.98°C for entire India for the same period.The Himalayan states are experiencing more severe impacts of global warming.The present investigation found a strong correlation (0.98)between the average temperature trend for all the Himalayan states.The correlation coefficient between temperature and rainfall was significantly strong for Northeastern Himalayan states A&M (0.80),WB&S (0.78), NMMT (0.76), and Arun (0.62); however, it was weak for Northwestern Himalayan states UK(0.5),Him(0.39)and J&K(0.18).

    The present investigation developed the CDLSTM model containing eight LSTM layers, where four layers were dense,and three were the dropout layers.The CDLSTM model was optimized based on rigorous parameters tuning.The developed CDLSTM model showed promising performance based on various metrics such as R2, MSE, RMSE, MAPE, MAD, and NSE.The developed CDLSTM model was likely to estimate the possible future values of temperature and rainfall accurately,given its reliability.The FB-Prophet model implemented in the present investigation with the PWL algorithm showed remarkably efficient performance based on accuracy metrics.As per available literature,the current investigation’s performance achieved by the FB-Prophet model for temperature and rainfall forecasting is the highest, based on accuracy metrics.The developed CDLSTM model has lower accuracy than the FB-Prophet model; however, the CDLSTM model showed better performance than other models applied in previous studies.Both CDLSTM and FB-Prophet model’s performance showed good forecasting values for all months, including Jan 2013, where the temperature was low due to the peak winter season.The future scope of the present investigation is to add more data on snow retreat,glacier melt,agricultural yield,and demographics to assess the complete cycle of climate change for the Himalayan region.Another future scope of the present investigation is to implements and assimilate the latest state of the art models for climate modeling and forecasting[30–34].

    6 Limitations and Learning Points of the Present Investigation

    The significant limitations of the present study include(1)Although the performance of the developed CDLSTM model was significantly higher than previous studies,the imported FB-Prophet model with PWL algorithm performed better than the developed CDLSTM model.(2) The computation of the tuned CDLSTM model took 15 minutes for 20 epochs, so an improvement in computational efficiency is required.(3)The reasons to choose LSTM in the present investigation are its capability to deal with the vanishing gradient problem and better control, flexibility, and performance than traditional RNN.(4) The LSTM model has limitations such as the requirement of high memory bandwidth due to linear layers;also,it is more prone to overfitting and is too complex to apply dropout,(5)The effect of Gulfstream weakening on climate change on agricultural productivity will be a future scope as parts of the US and Europe are influenced by the Gulf Stream.

    Acknowledgement:The author is thankful to IMD and the University of Berkeley for providing climate datasets.

    Funding Statement:Mohd Anul Haq would like to thank Deanship of Scientific Research at Majmaah University for supporting this work under Project No.R-2021-236.

    Conflicts of Interest:The author declare that they have no conflicts of interest to report regarding the present study.

    狠狠狠狠99中文字幕| 村上凉子中文字幕在线| 国产主播在线观看一区二区| 亚洲国产色片| 久久国产乱子伦精品免费另类| 国产久久久一区二区三区| 亚洲av成人av| 国产极品精品免费视频能看的| 精品人妻偷拍中文字幕| 一级作爱视频免费观看| 亚洲av.av天堂| 国产精品98久久久久久宅男小说| 国产综合懂色| 国产成人av教育| 欧美成人一区二区免费高清观看| 嫩草影院精品99| 亚洲精品亚洲一区二区| netflix在线观看网站| 无人区码免费观看不卡| 免费大片18禁| 日韩欧美 国产精品| 一夜夜www| 色噜噜av男人的天堂激情| av福利片在线观看| 美女被艹到高潮喷水动态| 非洲黑人性xxxx精品又粗又长| 亚洲无线在线观看| 国内揄拍国产精品人妻在线| 欧美不卡视频在线免费观看| 久久国产精品人妻蜜桃| 日本 欧美在线| 91在线观看av| 一边摸一边抽搐一进一小说| 国内精品久久久久精免费| 午夜免费成人在线视频| 搡老妇女老女人老熟妇| 亚洲国产色片| 日韩精品中文字幕看吧| 国产精品一区二区三区四区久久| 丰满乱子伦码专区| 国产成年人精品一区二区| 91麻豆av在线| 国内精品美女久久久久久| а√天堂www在线а√下载| 中文字幕av在线有码专区| 日本一二三区视频观看| 美女 人体艺术 gogo| 精品免费久久久久久久清纯| 欧美一区二区精品小视频在线| 久久久久性生活片| 亚洲欧美日韩无卡精品| 国产精品乱码一区二三区的特点| 亚洲第一欧美日韩一区二区三区| 有码 亚洲区| 午夜福利高清视频| 亚洲不卡免费看| 少妇人妻精品综合一区二区 | 国产精品1区2区在线观看.| 好男人在线观看高清免费视频| av国产免费在线观看| 人妻制服诱惑在线中文字幕| 中文字幕精品亚洲无线码一区| 免费观看人在逋| 国产av在哪里看| 老司机深夜福利视频在线观看| 人妻久久中文字幕网| 日本免费a在线| 97人妻精品一区二区三区麻豆| 人人妻人人看人人澡| x7x7x7水蜜桃| 亚洲黑人精品在线| 亚洲午夜理论影院| 在线看三级毛片| 国产成+人综合+亚洲专区| 淫妇啪啪啪对白视频| 午夜免费激情av| 国内毛片毛片毛片毛片毛片| 国产精品三级大全| 每晚都被弄得嗷嗷叫到高潮| 麻豆一二三区av精品| 日韩有码中文字幕| 久久久久久久精品吃奶| 久久久久九九精品影院| 午夜两性在线视频| 非洲黑人性xxxx精品又粗又长| 亚洲精品成人久久久久久| 色av中文字幕| 色综合亚洲欧美另类图片| 国产色婷婷99| 国产成人福利小说| 久久人妻av系列| 99在线人妻在线中文字幕| 亚洲欧美日韩无卡精品| 国产黄片美女视频| 国产高清三级在线| 亚洲aⅴ乱码一区二区在线播放| 丰满的人妻完整版| 亚洲一区二区三区色噜噜| 午夜精品一区二区三区免费看| 熟女人妻精品中文字幕| 成人午夜高清在线视频| 色噜噜av男人的天堂激情| 午夜两性在线视频| 国产免费男女视频| 99热精品在线国产| 欧美极品一区二区三区四区| 真人一进一出gif抽搐免费| 舔av片在线| 热99在线观看视频| 在线观看免费视频日本深夜| 国产又黄又爽又无遮挡在线| 亚洲狠狠婷婷综合久久图片| 两个人的视频大全免费| 成人精品一区二区免费| 精品人妻熟女av久视频| 亚洲av熟女| 国产精品久久电影中文字幕| 欧美另类亚洲清纯唯美| 在线观看美女被高潮喷水网站 | 91久久精品电影网| 精品久久久久久久人妻蜜臀av| 欧美成人免费av一区二区三区| 老鸭窝网址在线观看| 琪琪午夜伦伦电影理论片6080| 欧美又色又爽又黄视频| 网址你懂的国产日韩在线| 午夜免费男女啪啪视频观看 | 特级一级黄色大片| 级片在线观看| 欧美中文日本在线观看视频| 99热只有精品国产| 看片在线看免费视频| 欧美最黄视频在线播放免费| 欧美极品一区二区三区四区| 能在线免费观看的黄片| 亚洲精品影视一区二区三区av| 我要看日韩黄色一级片| 国产精品av视频在线免费观看| 国产男靠女视频免费网站| 欧美日本亚洲视频在线播放| 97碰自拍视频| 国产高清视频在线播放一区| 在线a可以看的网站| 久久久久久久久久成人| 真人做人爱边吃奶动态| 久久久久久久久久黄片| av欧美777| 国内揄拍国产精品人妻在线| 日本 av在线| 综合色av麻豆| 99热精品在线国产| 91在线观看av| 99久久精品一区二区三区| 日本三级黄在线观看| 国产精品久久电影中文字幕| 日本在线视频免费播放| 久久草成人影院| 国产高清有码在线观看视频| 特级一级黄色大片| 日本 欧美在线| 亚洲狠狠婷婷综合久久图片| 欧美一区二区国产精品久久精品| 一二三四社区在线视频社区8| 国产精品免费一区二区三区在线| 国产91精品成人一区二区三区| 国产亚洲欧美98| 美女cb高潮喷水在线观看| 老鸭窝网址在线观看| 91在线精品国自产拍蜜月| 欧美日韩乱码在线| 国产精品,欧美在线| 国产精品综合久久久久久久免费| 色5月婷婷丁香| 亚洲人成网站高清观看| 一个人免费在线观看电影| 欧美精品国产亚洲| av欧美777| 欧美+亚洲+日韩+国产| 国产一区二区三区视频了| 99久久99久久久精品蜜桃| 内地一区二区视频在线| 在线观看午夜福利视频| 国产一区二区在线av高清观看| 中文字幕av成人在线电影| 日韩精品中文字幕看吧| 99国产极品粉嫩在线观看| 成人午夜高清在线视频| 久久久成人免费电影| 99视频精品全部免费 在线| 好看av亚洲va欧美ⅴa在| 老司机午夜十八禁免费视频| 十八禁人妻一区二区| 天天躁日日操中文字幕| 色哟哟·www| 丰满的人妻完整版| 国产精品综合久久久久久久免费| 精品福利观看| 18+在线观看网站| 精品无人区乱码1区二区| 成人av在线播放网站| 一区二区三区四区激情视频 | 琪琪午夜伦伦电影理论片6080| 自拍偷自拍亚洲精品老妇| 亚洲色图av天堂| 国产大屁股一区二区在线视频| 色尼玛亚洲综合影院| 国产精品美女特级片免费视频播放器| 成人特级av手机在线观看| 免费av毛片视频| 一级a爱片免费观看的视频| 亚洲一区二区三区色噜噜| 亚洲一区二区三区不卡视频| 久久人妻av系列| 免费黄网站久久成人精品 | 中文字幕av在线有码专区| 色av中文字幕| 欧美色欧美亚洲另类二区| 精品日产1卡2卡| 国产精品一区二区免费欧美| 少妇的逼水好多| 国产精品野战在线观看| 国产男靠女视频免费网站| 欧美一区二区国产精品久久精品| 最新中文字幕久久久久| 大型黄色视频在线免费观看| 日韩精品中文字幕看吧| 国产探花极品一区二区| 成人美女网站在线观看视频| 日日干狠狠操夜夜爽| 国产一区二区亚洲精品在线观看| 亚洲内射少妇av| 久久婷婷人人爽人人干人人爱| 婷婷精品国产亚洲av在线| 91久久精品电影网| 免费看a级黄色片| 一区二区三区高清视频在线| 亚洲第一区二区三区不卡| 熟女电影av网| 99国产精品一区二区三区| 久久久久久国产a免费观看| 亚洲熟妇中文字幕五十中出| 97人妻精品一区二区三区麻豆| 国产av在哪里看| 色综合亚洲欧美另类图片| 亚洲国产精品sss在线观看| 人人妻人人澡欧美一区二区| 91午夜精品亚洲一区二区三区 | 搡老岳熟女国产| 亚洲va日本ⅴa欧美va伊人久久| 又黄又爽又免费观看的视频| 国产精品电影一区二区三区| 国产探花在线观看一区二区| 亚洲无线观看免费| 国产精品98久久久久久宅男小说| 有码 亚洲区| 精华霜和精华液先用哪个| 国产毛片a区久久久久| 老熟妇仑乱视频hdxx| 欧美+日韩+精品| 免费一级毛片在线播放高清视频| 亚洲男人的天堂狠狠| 国产aⅴ精品一区二区三区波| 美女高潮的动态| 97超级碰碰碰精品色视频在线观看| 夜夜爽天天搞| 一区二区三区四区激情视频 | 俺也久久电影网| 亚洲av第一区精品v没综合| 免费搜索国产男女视频| 天堂av国产一区二区熟女人妻| 欧美一级a爱片免费观看看| 全区人妻精品视频| 国内精品美女久久久久久| 在线观看av片永久免费下载| 国产三级中文精品| 国产伦精品一区二区三区视频9| 国产av不卡久久| 69av精品久久久久久| 亚洲人与动物交配视频| 亚洲美女搞黄在线观看 | 国产激情偷乱视频一区二区| 亚洲av成人精品一区久久| 成人精品一区二区免费| 亚洲,欧美,日韩| 永久网站在线| 中亚洲国语对白在线视频| 亚洲第一区二区三区不卡| 精品不卡国产一区二区三区| 精品一区二区三区人妻视频| 亚洲精品一区av在线观看| 日韩中字成人| 五月伊人婷婷丁香| av在线蜜桃| 婷婷亚洲欧美| 网址你懂的国产日韩在线| 国产精品99久久久久久久久| 熟妇人妻久久中文字幕3abv| 少妇熟女aⅴ在线视频| 亚洲一区二区三区不卡视频| 国产毛片a区久久久久| 九色国产91popny在线| 欧美色欧美亚洲另类二区| 91在线精品国自产拍蜜月| 久久人人爽人人爽人人片va | 在线观看免费视频日本深夜| 内射极品少妇av片p| 国产乱人伦免费视频| av在线蜜桃| av天堂在线播放| 国产精品影院久久| 午夜影院日韩av| 午夜影院日韩av| 两个人的视频大全免费| 亚洲精品日韩av片在线观看| 亚洲成人免费电影在线观看| 永久网站在线| 欧美日本亚洲视频在线播放| 一二三四社区在线视频社区8| 12—13女人毛片做爰片一| 亚洲 国产 在线| 一区福利在线观看| 亚州av有码| 中亚洲国语对白在线视频| 欧美另类亚洲清纯唯美| 久久精品国产亚洲av涩爱 | 免费在线观看成人毛片| 精品久久久久久久末码| 一区二区三区免费毛片| 99久久无色码亚洲精品果冻| 亚洲欧美日韩高清专用| 免费看日本二区| 国产麻豆成人av免费视频| 国产精品伦人一区二区| 成年免费大片在线观看| 在线播放国产精品三级| 蜜桃久久精品国产亚洲av| 看免费av毛片| 久久精品国产亚洲av涩爱 | 成熟少妇高潮喷水视频| 成人一区二区视频在线观看| 日本黄色视频三级网站网址| 香蕉av资源在线| 欧美午夜高清在线| 自拍偷自拍亚洲精品老妇| 伊人久久精品亚洲午夜| 99久久精品国产亚洲精品| 在现免费观看毛片| 一二三四社区在线视频社区8| 国产av一区在线观看免费| 欧美精品国产亚洲| 亚洲无线在线观看| 国产黄色小视频在线观看| 一本久久中文字幕| 国产单亲对白刺激| 国产精品一区二区免费欧美| 国产免费男女视频| 最好的美女福利视频网| 69人妻影院| 十八禁网站免费在线| 国产精品自产拍在线观看55亚洲| 免费人成视频x8x8入口观看| 好男人在线观看高清免费视频| 成年免费大片在线观看| 99热这里只有是精品50| 91九色精品人成在线观看| h日本视频在线播放| 欧美乱色亚洲激情| 夜夜看夜夜爽夜夜摸| 欧美+亚洲+日韩+国产| 免费电影在线观看免费观看| 美女xxoo啪啪120秒动态图 | 亚洲av一区综合| 岛国在线免费视频观看| 午夜福利视频1000在线观看| 最后的刺客免费高清国语| 69人妻影院| 国产日本99.免费观看| 午夜激情欧美在线| 国产av一区在线观看免费| 在线国产一区二区在线| 亚洲av二区三区四区| 欧美日本亚洲视频在线播放| 国产高清激情床上av| 露出奶头的视频| 精品一区二区三区视频在线| 免费黄网站久久成人精品 | 国产熟女xx| 每晚都被弄得嗷嗷叫到高潮| 国产免费av片在线观看野外av| 最近视频中文字幕2019在线8| 啪啪无遮挡十八禁网站| 国产精华一区二区三区| 中文字幕免费在线视频6| 99久久99久久久精品蜜桃| 日韩欧美 国产精品| 国产不卡一卡二| 一级毛片久久久久久久久女| 亚洲三级黄色毛片| 日韩欧美一区二区三区在线观看| 午夜激情福利司机影院| 国产毛片a区久久久久| 黄色女人牲交| 毛片女人毛片| 好男人在线观看高清免费视频| 一卡2卡三卡四卡精品乱码亚洲| 日韩欧美在线乱码| 精品人妻偷拍中文字幕| 他把我摸到了高潮在线观看| 淫秽高清视频在线观看| 一本一本综合久久| 天堂影院成人在线观看| 国产探花极品一区二区| 亚洲真实伦在线观看| 久久久久久国产a免费观看| 可以在线观看的亚洲视频| 人人妻人人看人人澡| 色哟哟哟哟哟哟| 国产伦精品一区二区三区视频9| 嫩草影院新地址| 国产精品伦人一区二区| 能在线免费观看的黄片| 级片在线观看| av天堂中文字幕网| 亚洲精品成人久久久久久| 色哟哟·www| 精品欧美国产一区二区三| 亚洲精品乱码久久久v下载方式| 欧美潮喷喷水| 不卡一级毛片| 亚州av有码| 亚洲精华国产精华精| 午夜福利成人在线免费观看| 亚洲午夜理论影院| 精品人妻1区二区| 亚洲av日韩精品久久久久久密| 亚洲国产色片| 国产精品人妻久久久久久| 91久久精品电影网| АⅤ资源中文在线天堂| 亚洲av第一区精品v没综合| 高清在线国产一区| 国产极品精品免费视频能看的| 少妇被粗大猛烈的视频| 乱人视频在线观看| 在线免费观看不下载黄p国产 | 午夜免费男女啪啪视频观看 | 亚洲欧美日韩卡通动漫| 午夜福利免费观看在线| 欧美日韩福利视频一区二区| 乱码一卡2卡4卡精品| 欧美+亚洲+日韩+国产| 国产精品,欧美在线| 成人av在线播放网站| 人妻丰满熟妇av一区二区三区| 国产精品久久久久久久电影| 亚洲无线观看免费| 三级国产精品欧美在线观看| 少妇被粗大猛烈的视频| 欧美成人a在线观看| 精品日产1卡2卡| 韩国av一区二区三区四区| ponron亚洲| 欧美精品国产亚洲| 国产成人啪精品午夜网站| 亚洲午夜理论影院| 最好的美女福利视频网| 我要看日韩黄色一级片| 国产69精品久久久久777片| 欧美日韩瑟瑟在线播放| 亚洲综合色惰| 尤物成人国产欧美一区二区三区| 国产精品综合久久久久久久免费| 露出奶头的视频| 国产日本99.免费观看| 欧美日本亚洲视频在线播放| 人妻夜夜爽99麻豆av| 亚洲av免费在线观看| 在现免费观看毛片| 久久久久久久亚洲中文字幕 | 草草在线视频免费看| 最近中文字幕高清免费大全6 | 国产在视频线在精品| 老熟妇乱子伦视频在线观看| 少妇丰满av| 国产精品精品国产色婷婷| 午夜激情欧美在线| 一本综合久久免费| 18禁黄网站禁片免费观看直播| 中文字幕熟女人妻在线| 国内精品久久久久久久电影| 免费在线观看日本一区| 婷婷精品国产亚洲av在线| 久久人人爽人人爽人人片va | 伦理电影大哥的女人| 每晚都被弄得嗷嗷叫到高潮| 国产野战对白在线观看| 免费高清视频大片| a级毛片免费高清观看在线播放| 天天躁日日操中文字幕| 禁无遮挡网站| 国产成人欧美在线观看| 日本一本二区三区精品| 午夜日韩欧美国产| 18+在线观看网站| 人人妻人人看人人澡| 亚洲va日本ⅴa欧美va伊人久久| 两个人的视频大全免费| 3wmmmm亚洲av在线观看| 亚洲av日韩精品久久久久久密| 欧美zozozo另类| x7x7x7水蜜桃| 国产免费男女视频| 男人舔奶头视频| 免费看日本二区| 757午夜福利合集在线观看| 欧美bdsm另类| 99国产精品一区二区蜜桃av| 国产精品亚洲av一区麻豆| 免费无遮挡裸体视频| 国产精品免费一区二区三区在线| 成人av在线播放网站| netflix在线观看网站| 国产精品一区二区三区四区久久| 国产精品一及| 久久精品综合一区二区三区| 成人性生交大片免费视频hd| 91狼人影院| www日本黄色视频网| 免费人成视频x8x8入口观看| 亚洲在线自拍视频| 波多野结衣高清无吗| 色播亚洲综合网| 人妻丰满熟妇av一区二区三区| 男人舔奶头视频| 啪啪无遮挡十八禁网站| 亚洲av成人精品一区久久| av在线观看视频网站免费| 欧美成狂野欧美在线观看| 日日摸夜夜添夜夜添av毛片 | 欧洲精品卡2卡3卡4卡5卡区| 成年版毛片免费区| 亚州av有码| 午夜激情欧美在线| 日韩欧美在线乱码| 国产不卡一卡二| 一本一本综合久久| 少妇的逼水好多| 久久精品久久久久久噜噜老黄 | 97碰自拍视频| 国产成人福利小说| 51国产日韩欧美| 欧美绝顶高潮抽搐喷水| 国产蜜桃级精品一区二区三区| 国产人妻一区二区三区在| 人妻制服诱惑在线中文字幕| 宅男免费午夜| 色视频www国产| 欧美区成人在线视频| 噜噜噜噜噜久久久久久91| 丰满人妻熟妇乱又伦精品不卡| 成人毛片a级毛片在线播放| 久久精品国产自在天天线| 午夜精品久久久久久毛片777| 国产精品乱码一区二三区的特点| 日韩欧美三级三区| 日本黄色片子视频| 麻豆成人av在线观看| 露出奶头的视频| 高清在线国产一区| 91字幕亚洲| 18美女黄网站色大片免费观看| 看免费av毛片| 丁香欧美五月| .国产精品久久| 天堂影院成人在线观看| 亚洲av美国av| 老师上课跳d突然被开到最大视频 久久午夜综合久久蜜桃 | 成人国产一区最新在线观看| 精品欧美国产一区二区三| 亚洲中文字幕一区二区三区有码在线看| 午夜福利视频1000在线观看| 两性午夜刺激爽爽歪歪视频在线观看| 亚洲国产欧美人成| 午夜视频国产福利| 亚洲av成人av| 一卡2卡三卡四卡精品乱码亚洲| 一区二区三区激情视频| 国产成人福利小说| 一进一出抽搐动态| 久久久久久久久久成人| www.色视频.com| 亚洲,欧美精品.| 日韩av在线大香蕉| 久久久国产成人精品二区| 少妇高潮的动态图| 国产伦精品一区二区三区视频9| 成人性生交大片免费视频hd| 黄色日韩在线| 观看美女的网站| 激情在线观看视频在线高清| 老司机深夜福利视频在线观看| 一a级毛片在线观看| 久久亚洲真实| 国产单亲对白刺激| 精品福利观看| 变态另类丝袜制服| 国产精品伦人一区二区| 久久欧美精品欧美久久欧美| 麻豆成人午夜福利视频| 一本综合久久免费| 久久精品夜夜夜夜夜久久蜜豆| 蜜桃久久精品国产亚洲av| 婷婷六月久久综合丁香| 国产精华一区二区三区| 亚洲色图av天堂|