SlideShare a Scribd company logo
1 of 24
Prediction of droughts over Pakistan using
machine learning algorithms
Presented By
KARTIK JADAV(19AG62R19)
Agricultural and Food Engineering Department, Indian Institute
of Technology Kharagpur, India
Reference : Najeebullah Khan, D.A. Sachindra , Shamsuddin Shahid , Kamal Ahmed a ,
Mohammed Sanusi Shiru , Nadeem Nawaz “Prediction of droughts over Pakistan using
machine learning algorithms ”, Advances in Water Resources(2020) Volume 139, May
2020, 103562
Motivation
 Why study had been done?
 The main objective of this study was to develop ML-based models for
predicting moderate, severe and extreme droughts over Pakistan based on
the SPEI.
 Context of the study
 In this study, the drought prediction models were developed using ANN,
SVM and KNN.
 Contribution
 As droughts have a direct relationship with the availability of water, their
changing characteristics due to changing climate will have a pro- found
impact on water stress and food security.
REVIEW OF LITERATURE
Author Year Findings
Ganguli and
Reddy
2014 SVM is able to learn from a much smaller data set and
is capable of handling a large number of variables. SVM
may overcome certain limitations of ANN such as
trapping at local minima and overfitting, to some extent
in predicting droughts
Yang et al 2015 Among the ML techniques, ANN and SVM can be
regarded as the most widely used techniques for
developing drought prediction models
Study area and datasets
• Pakistan located in South
Asia covers an area of
796,095 km2.
• Pakistan experiences two
monsoon precipitation
seasons; IM (July–Sep),
WD (Dec–Mar).
Fig. 1. IM refers to the regions which receive
precipitation mainly during the Indian monsoon and
WD refers to the regions which receive
precipitation mainly during the western
disturbance.
• It experiences four seasons (based on temperature) cool and dry winter
(Dec–Feb), hot and dry spring (Mar–May), hot and humid summer (Jun–Aug)
and dry autumn (Sep–Nov)
• During these two precipitation seasons, two important cropping seasons
called Kharif (May-Oct) observed during the Indian monsoon period
• The Rabi (Nov-Apr) is experienced during the western disturbances
• These cropping seasons play an important role in agriculture and the
economy of Pakistan
Gridded precipitation and
temperature datasets
• This study uses the Princeton global forcing (PGF) gridded temperature
and precipitation datasets for estimating the SPEI (NCEP/NCAR datasets)
• The PGF gridded temperature and precipitation datasets have a resolution
of 0.25 ×0.25 in the latitudinal and longitudinal directions and hence the
data are available at 1437 grid points over Pakistan.
• These data were used in the calculation of SPEI at each of the 1437 grid
points distributed over Pakistan, Among which 531 grid points lie in the
region of WD and 906 grid points lie in the IM dominated region
Method
1–6-month SPEIs were calculated for each of the grid points
Based on 1–6-month SPEIs droughts were categorized
Different large-scale atmospheric variables (probable predictors) that are
influential on droughts were identified based on past literature
Principle Component Analysis (PCA) was used to generate the Principal
Components (PCs) from the data
The difference between the PCs for February and April(Kharif) and difference
between the PCs for October and August(Rabi) were computed.
The PCs selected using SVM-RFE they were used in the development of
drought prediction models
the correlations between the PCs and 1–6-month SPEI were separately
calculated
ML-algorithms were used to develop drought prediction
PCA
Why?
• Overfitting (Due to too much features and attribute we gave
during training phase so model is confused and problem of
overfitting occur)
• PCA is Linear projection method to reduce number of
parameters(Attributes and features).
1. Standardize the data
2. Calculate the covariance matrix
3. Find the eigenvalues and eigenvectors of covariance matrix
4. Plot the eigenvectors/principal components over scaled data
SVM
• SVM is a widely used ML algorithm that can be used in developing
classification and regression models.
• However, Sachindra et al. (2018) showed that the polynomial kernel
performs better with SVM. Hence, the polynomial kernel was used in this
study.
• The polynomial kernel is defined as
. 𝐾 (𝑥 𝑖 , 𝑥 𝑗 )= (𝑥 𝑖 . 𝑥 𝑗 + 𝑐 ) 𝑑
Where, 𝑥 𝑖 and 𝑥 𝑗 are the predictor and predictand data, d is the degree of
the polynomial, and c is a constant that allows a trade-off between the
influence of the higher and lower order terms.
ANN
• .To reduce the chances of overfitting, a technique known as Bayesian
regularization was developed ( Burden and Winkler, 2008 ; MacKay,
1992 ; Ticknor, 2013 ).
𝐹 = 𝛽E D + 𝛼E w
• In the Bayesian network, the weights are considered random variables
and thus their density function is expressed following the Bayes’ rules
𝑃 ( 𝑤 |𝐷, 𝛼, 𝛽, M) =
𝑃 (D |w , 𝛽, M) P(w |𝛼, M)
𝑃 (D |𝛼, 𝛽, M)
Where w is the vector of network weights, D
represents the predictor and the predictand data
vectors (x, y), and M is the neural network model
being used, 𝛼 and 𝛽 are Regularization
K-Nearest Neighbour
• The KNN for classification and regression is the simplest non-parametric ML
techniques
• KNN algorithm finds the k Nearest Neighbours from calibration data using a
distance measure such as the Euclidean distance
𝑑 𝑗𝑜 = 𝑖=1
݊ (𝑥 𝑖𝑗 − 𝑥 𝑖𝑜 )2 𝑡 = 1 , 2 , 3 , …., ݊
• For calculating KNN simulated values ( Zr ) Zk refers to the neighbouring data
and f k ( dj ) is the kernel function
𝑍 𝑟 = ݇ =1
𝐾 ݂݇ (𝑑𝑗 ×𝑍݇ ).
Performance assessment
1) The coefficient of determination (R2 )
varies between 0 (no agreement) and 1
(perfect agreement)
2) The normalized root mean squared error
(NRMSE) varies between 0 (perfect
agreement) and +∞(no agreement)
3) Percentage of bias (Pbias) varies between
- ∞(underestimation) and +∞(over-
estimation) and a value of 0 refers to a
perfect agreement
4) lThe modified index of agreement ( md )
varies between 0 (no agreement) and 1
(perfect agreement)
N is the number of testing samples. Oi are
the ith observation, Si are the i th simulated
value, sd is the standard deviation of the
observations.
Results and Discussion
Fig. 2. Climate domain used in the present study, 861 NCEP/NCAR grid
points spread across the climate domain are depicted in grey circles.
Fig. 3. The spatial pattern of the correlation coefficient of the RH, atm temperature and wind
component (925hPa) during the Rabi and Kharif periods. The correlation is calculated with
6-month SPEI in April for Rabi and in October for Kharif. The dots represent the significant
correlation at 95% confidence level with the SPEI.
Fig. 4. Boxplots showing correlations of the selected atmospheric variables at four
pressure levels i.e. 925, 850, 700 and 500 hPa (except SLP) with the SPEI of Pakistan
Red boxplots (Rabi season) Blue (Kharif season)
Fig. 5. Performance of different machine learning-based models in
predicting moderate droughts in the Rabi season.
Fig. 6. Performance of different machine learning-based models
in predicting moderate droughts in the Kharif season.
Fig.7. Boxplots of the spatial correlation and NRMSE for the Rabi and Kharif
seasons for the month of April shown in red and blue colours respectively.
Conclusions
• In this study, it was seen that KNN-based drought models display limited
performance comparison to other drought models
• SVM-based models were able to better capture the temporal and spatial
characteristics of droughts
• It was found that in the Rabi season SPEI is positively correlated with RH
over the Mediterranean Sea and the region north of the Caspian Sea,
whereas in the Kharif season SPEI is positively correlated with the humid
region over the south-eastern part of the Bay of Bengal and Caspian Seas.
• Rabi season SPEI also displayed statistically significant positive
correlations with atmospheric temperature over the region north of the
Mediterranean Sea.
• Also, wind speeds over the Indian Ocean and the Arabian Sea during
the Kharif season were statistically significantly correlated with SPEI
.
• This indicated that RH, temperature and wind speed are indicators of
droughts over Pakistan.
Limitations of this study
• The present study used only one reanalysis dataset (NCEP/NCAR) to provide
inputs to the drought prediction models.
• In order to study the uncertainties which originate from the inputs, different
reanalysis datasets should be used in developing drought prediction models.
• However, this was not performed as the other major reanalysis datasets
either did not have data for some of the specific atmospheric variables used
in this study or they did not have data corresponding to the period 1948–
2016.
Future scope
• Sachindra and Kanae (2019) proved that Parallel Multi-Population Genetic
Programming (PMPGP)-based statistical downscaling models show better
generalization skills and higher resistance to redundant information in
inputs compared to those based on traditional GP.
• Thus, other ML algorithms such as Extreme Learning Machine (ELM),
Genetic Programming (GP), PMPGP, and Random Forest (RF) should also
be tested in future for developing drought prediction models over Pakistan.
References
• Najeebullah Khan, D.A. Sachindra , Shamsuddin Shahid ,
Kamal Ahmed a , Mohammed Sanusi Shiru , Nadeem Nawaz
“Prediction of droughts over Pakistan using machine learning
algorithms ”, Advances in Water Resources(2020) Volume
139, May 2020, 103562
• Ganguli, P. , Reddy, M.J. , 2014. Ensemble prediction of
regional droughts using climate inputs and the SVM–copula
approach. Hydrol. Process. 28, 4989–5009
Prediction droughts using machine learning algorithm

More Related Content

What's hot

Statistical downscaling sdsm
Statistical downscaling sdsmStatistical downscaling sdsm
Statistical downscaling sdsmPutu Santikayasa
 
Climate downscaling
Climate downscalingClimate downscaling
Climate downscalingIC3Climate
 
Climate change impact assessment on hydrology on river basins
Climate change impact assessment on hydrology on river basinsClimate change impact assessment on hydrology on river basins
Climate change impact assessment on hydrology on river basinsAbhiram Kanigolla
 
Groundwater modelling (an Introduction)
Groundwater modelling (an Introduction)Groundwater modelling (an Introduction)
Groundwater modelling (an Introduction)Putika Ashfar Khoiri
 
Drought and its management
Drought and its managementDrought and its management
Drought and its managementEshwarPrasad25
 
Flood risk assessment methodology
Flood risk assessment methodologyFlood risk assessment methodology
Flood risk assessment methodologyAmitSaha123
 
Drought indices
Drought indices Drought indices
Drought indices Rahul5110
 
Automatic weather station their working principle and importance
Automatic weather station their working principle and importanceAutomatic weather station their working principle and importance
Automatic weather station their working principle and importanceKhileshKumarsahu
 
Goswami Climate Change And Indian Monsoon Cse Workshop
Goswami  Climate Change And Indian Monsoon Cse WorkshopGoswami  Climate Change And Indian Monsoon Cse Workshop
Goswami Climate Change And Indian Monsoon Cse Workshopequitywatch
 
Modeling the Climate System: Is model-based science like model-based engineer...
Modeling the Climate System: Is model-based science like model-based engineer...Modeling the Climate System: Is model-based science like model-based engineer...
Modeling the Climate System: Is model-based science like model-based engineer...Steve Easterbrook
 
Effects of Climate change on water resources
Effects of Climate change on water resourcesEffects of Climate change on water resources
Effects of Climate change on water resourcesNjorBenedict1
 
Floodmanagementstrategies
FloodmanagementstrategiesFloodmanagementstrategies
FloodmanagementstrategiesWill Williams
 
Climate Change Impact Assessment on Hydrological Regime of Kali Gandaki Basin
Climate Change Impact Assessment on Hydrological Regime of Kali Gandaki BasinClimate Change Impact Assessment on Hydrological Regime of Kali Gandaki Basin
Climate Change Impact Assessment on Hydrological Regime of Kali Gandaki BasinHI-AWARE
 
REMOTE SENSING DATA FOR HYDROLOGICAL MODELING
REMOTE SENSING DATA FOR HYDROLOGICAL MODELINGREMOTE SENSING DATA FOR HYDROLOGICAL MODELING
REMOTE SENSING DATA FOR HYDROLOGICAL MODELINGShyam Mohan Chaudhary
 
Advances in agricultural drought monitoring and forecasting
Advances in agricultural drought monitoring and forecastingAdvances in agricultural drought monitoring and forecasting
Advances in agricultural drought monitoring and forecastingAbhilash Singh Chauhan
 

What's hot (20)

Statistical downscaling sdsm
Statistical downscaling sdsmStatistical downscaling sdsm
Statistical downscaling sdsm
 
Swat & modflow
Swat & modflowSwat & modflow
Swat & modflow
 
Climate downscaling
Climate downscalingClimate downscaling
Climate downscaling
 
Climate change impact assessment on hydrology on river basins
Climate change impact assessment on hydrology on river basinsClimate change impact assessment on hydrology on river basins
Climate change impact assessment on hydrology on river basins
 
Groundwater modelling (an Introduction)
Groundwater modelling (an Introduction)Groundwater modelling (an Introduction)
Groundwater modelling (an Introduction)
 
Drought and its management
Drought and its managementDrought and its management
Drought and its management
 
Introduction to flood modelling
Introduction to flood modellingIntroduction to flood modelling
Introduction to flood modelling
 
Flood risk assessment methodology
Flood risk assessment methodologyFlood risk assessment methodology
Flood risk assessment methodology
 
Drought indices
Drought indices Drought indices
Drought indices
 
Automatic weather station their working principle and importance
Automatic weather station their working principle and importanceAutomatic weather station their working principle and importance
Automatic weather station their working principle and importance
 
Climate and climate modelling
Climate and climate modellingClimate and climate modelling
Climate and climate modelling
 
Goswami Climate Change And Indian Monsoon Cse Workshop
Goswami  Climate Change And Indian Monsoon Cse WorkshopGoswami  Climate Change And Indian Monsoon Cse Workshop
Goswami Climate Change And Indian Monsoon Cse Workshop
 
Modeling the Climate System: Is model-based science like model-based engineer...
Modeling the Climate System: Is model-based science like model-based engineer...Modeling the Climate System: Is model-based science like model-based engineer...
Modeling the Climate System: Is model-based science like model-based engineer...
 
DROUGHT INDEX
DROUGHT INDEXDROUGHT INDEX
DROUGHT INDEX
 
Effects of Climate change on water resources
Effects of Climate change on water resourcesEffects of Climate change on water resources
Effects of Climate change on water resources
 
Floodmanagementstrategies
FloodmanagementstrategiesFloodmanagementstrategies
Floodmanagementstrategies
 
Climate Change Impact Assessment on Hydrological Regime of Kali Gandaki Basin
Climate Change Impact Assessment on Hydrological Regime of Kali Gandaki BasinClimate Change Impact Assessment on Hydrological Regime of Kali Gandaki Basin
Climate Change Impact Assessment on Hydrological Regime of Kali Gandaki Basin
 
Swat model
Swat model Swat model
Swat model
 
REMOTE SENSING DATA FOR HYDROLOGICAL MODELING
REMOTE SENSING DATA FOR HYDROLOGICAL MODELINGREMOTE SENSING DATA FOR HYDROLOGICAL MODELING
REMOTE SENSING DATA FOR HYDROLOGICAL MODELING
 
Advances in agricultural drought monitoring and forecasting
Advances in agricultural drought monitoring and forecastingAdvances in agricultural drought monitoring and forecasting
Advances in agricultural drought monitoring and forecasting
 

Similar to Prediction droughts using machine learning algorithm

Time Series Data Analysis for Forecasting – A Literature Review
Time Series Data Analysis for Forecasting – A Literature ReviewTime Series Data Analysis for Forecasting – A Literature Review
Time Series Data Analysis for Forecasting – A Literature ReviewIJMER
 
Application of mathematical modelling in rainfall forcast a csae study in...
Application of mathematical modelling in rainfall forcast     a csae study in...Application of mathematical modelling in rainfall forcast     a csae study in...
Application of mathematical modelling in rainfall forcast a csae study in...eSAT Journals
 
Assessment of two Methods to study Precipitation Prediction
Assessment of two Methods to study Precipitation PredictionAssessment of two Methods to study Precipitation Prediction
Assessment of two Methods to study Precipitation PredictionAI Publications
 
Application of the extreme learning machine algorithm for the
Application of the extreme learning machine algorithm for theApplication of the extreme learning machine algorithm for the
Application of the extreme learning machine algorithm for themehmet şahin
 
IRJET- Rainfall Forecasting using Regression Techniques
IRJET- Rainfall Forecasting using Regression TechniquesIRJET- Rainfall Forecasting using Regression Techniques
IRJET- Rainfall Forecasting using Regression TechniquesIRJET Journal
 
Calculation of solar radiation by using regression methods
Calculation of solar radiation by using regression methodsCalculation of solar radiation by using regression methods
Calculation of solar radiation by using regression methodsmehmet şahin
 
A comparative study of different imputation methods for daily rainfall data i...
A comparative study of different imputation methods for daily rainfall data i...A comparative study of different imputation methods for daily rainfall data i...
A comparative study of different imputation methods for daily rainfall data i...journalBEEI
 
Comparison of ann and mlr models for
Comparison of ann and mlr models forComparison of ann and mlr models for
Comparison of ann and mlr models formehmet şahin
 
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...ijcseit
 
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...ijcseit
 
Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...Kaja Bantha Navas Raja Mohamed
 
Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...Kaja Bantha Navas Raja Mohamed
 
Estimation of precipitation during the period of south west monsoon
Estimation of precipitation during the period of south west monsoonEstimation of precipitation during the period of south west monsoon
Estimation of precipitation during the period of south west monsoonIAEME Publication
 
Binary classification of rainfall time-series using machine learning algorithms
Binary classification of rainfall time-series using machine  learning algorithmsBinary classification of rainfall time-series using machine  learning algorithms
Binary classification of rainfall time-series using machine learning algorithmsIJECEIAES
 
K venkata reddy
K venkata reddyK venkata reddy
K venkata reddyClimDev15
 
Downscaling global climate model outputs to fine scales over sri lanka for as...
Downscaling global climate model outputs to fine scales over sri lanka for as...Downscaling global climate model outputs to fine scales over sri lanka for as...
Downscaling global climate model outputs to fine scales over sri lanka for as...Pixel Clear (Pvt) Ltd
 
Comparative Study of Machine Learning Algorithms for Rainfall Prediction
Comparative Study of Machine Learning Algorithms for Rainfall PredictionComparative Study of Machine Learning Algorithms for Rainfall Prediction
Comparative Study of Machine Learning Algorithms for Rainfall Predictionijtsrd
 
Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...
Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...
Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...ijtsrd
 
RAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEY
RAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEYRAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEY
RAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEYcscpconf
 

Similar to Prediction droughts using machine learning algorithm (20)

Time Series Data Analysis for Forecasting – A Literature Review
Time Series Data Analysis for Forecasting – A Literature ReviewTime Series Data Analysis for Forecasting – A Literature Review
Time Series Data Analysis for Forecasting – A Literature Review
 
Application of mathematical modelling in rainfall forcast a csae study in...
Application of mathematical modelling in rainfall forcast     a csae study in...Application of mathematical modelling in rainfall forcast     a csae study in...
Application of mathematical modelling in rainfall forcast a csae study in...
 
Assessment of two Methods to study Precipitation Prediction
Assessment of two Methods to study Precipitation PredictionAssessment of two Methods to study Precipitation Prediction
Assessment of two Methods to study Precipitation Prediction
 
Application of the extreme learning machine algorithm for the
Application of the extreme learning machine algorithm for theApplication of the extreme learning machine algorithm for the
Application of the extreme learning machine algorithm for the
 
IRJET- Rainfall Forecasting using Regression Techniques
IRJET- Rainfall Forecasting using Regression TechniquesIRJET- Rainfall Forecasting using Regression Techniques
IRJET- Rainfall Forecasting using Regression Techniques
 
Calculation of solar radiation by using regression methods
Calculation of solar radiation by using regression methodsCalculation of solar radiation by using regression methods
Calculation of solar radiation by using regression methods
 
A comparative study of different imputation methods for daily rainfall data i...
A comparative study of different imputation methods for daily rainfall data i...A comparative study of different imputation methods for daily rainfall data i...
A comparative study of different imputation methods for daily rainfall data i...
 
APCBEE
APCBEEAPCBEE
APCBEE
 
Comparison of ann and mlr models for
Comparison of ann and mlr models forComparison of ann and mlr models for
Comparison of ann and mlr models for
 
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
 
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
FEED FORWARD BACK PROPAGATION NEURAL NETWORK COUPLED WITH RICE DATA SIMULATOR...
 
Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...
 
Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...Artificial Neural Network based computing model for wind speed prediction: A ...
Artificial Neural Network based computing model for wind speed prediction: A ...
 
Estimation of precipitation during the period of south west monsoon
Estimation of precipitation during the period of south west monsoonEstimation of precipitation during the period of south west monsoon
Estimation of precipitation during the period of south west monsoon
 
Binary classification of rainfall time-series using machine learning algorithms
Binary classification of rainfall time-series using machine  learning algorithmsBinary classification of rainfall time-series using machine  learning algorithms
Binary classification of rainfall time-series using machine learning algorithms
 
K venkata reddy
K venkata reddyK venkata reddy
K venkata reddy
 
Downscaling global climate model outputs to fine scales over sri lanka for as...
Downscaling global climate model outputs to fine scales over sri lanka for as...Downscaling global climate model outputs to fine scales over sri lanka for as...
Downscaling global climate model outputs to fine scales over sri lanka for as...
 
Comparative Study of Machine Learning Algorithms for Rainfall Prediction
Comparative Study of Machine Learning Algorithms for Rainfall PredictionComparative Study of Machine Learning Algorithms for Rainfall Prediction
Comparative Study of Machine Learning Algorithms for Rainfall Prediction
 
Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...
Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...
Runoff Prediction of Gharni River Catchment of Maharashtra by Regressional An...
 
RAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEY
RAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEYRAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEY
RAINFALL PREDICTION USING DATA MINING TECHNIQUES - A SURVEY
 

Recently uploaded

(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Serviceranjana rawat
 
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...Dr.Costas Sachpazis
 
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCollege Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCall Girls in Nagpur High Profile
 
IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024Mark Billinghurst
 
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur EscortsHigh Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escortsranjana rawat
 
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...ranjana rawat
 
ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...
ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...
ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...ZTE
 
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IVHARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IVRajaP95
 
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube ExchangerStudy on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube ExchangerAnamika Sarkar
 
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur High Profile
 
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINEMANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINESIVASHANKAR N
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Dr.Costas Sachpazis
 
Introduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptxIntroduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptxupamatechverse
 
Analog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog ConverterAnalog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog ConverterAbhinavSharma374939
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)Suman Mia
 
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escortsranjana rawat
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )Tsuyoshi Horigome
 
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptxBiology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptxDeepakSakkari2
 

Recently uploaded (20)

(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
 
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
Structural Analysis and Design of Foundations: A Comprehensive Handbook for S...
 
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCollege Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
 
IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024IVE Industry Focused Event - Defence Sector 2024
IVE Industry Focused Event - Defence Sector 2024
 
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur EscortsHigh Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
High Profile Call Girls Nagpur Isha Call 7001035870 Meet With Nagpur Escorts
 
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
 
ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...
ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...
ZXCTN 5804 / ZTE PTN / ZTE POTN / ZTE 5804 PTN / ZTE POTN 5804 ( 100/200 GE Z...
 
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCRCall Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
Call Us -/9953056974- Call Girls In Vikaspuri-/- Delhi NCR
 
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IVHARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
 
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube ExchangerStudy on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
 
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
 
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINEMANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
 
Introduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptxIntroduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptx
 
Analog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog ConverterAnalog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog Converter
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
 
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )
 
Biology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptxBiology for Computer Engineers Course Handout.pptx
Biology for Computer Engineers Course Handout.pptx
 
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
 

Prediction droughts using machine learning algorithm

  • 1. Prediction of droughts over Pakistan using machine learning algorithms Presented By KARTIK JADAV(19AG62R19) Agricultural and Food Engineering Department, Indian Institute of Technology Kharagpur, India Reference : Najeebullah Khan, D.A. Sachindra , Shamsuddin Shahid , Kamal Ahmed a , Mohammed Sanusi Shiru , Nadeem Nawaz “Prediction of droughts over Pakistan using machine learning algorithms ”, Advances in Water Resources(2020) Volume 139, May 2020, 103562
  • 2. Motivation  Why study had been done?  The main objective of this study was to develop ML-based models for predicting moderate, severe and extreme droughts over Pakistan based on the SPEI.  Context of the study  In this study, the drought prediction models were developed using ANN, SVM and KNN.  Contribution  As droughts have a direct relationship with the availability of water, their changing characteristics due to changing climate will have a pro- found impact on water stress and food security.
  • 3. REVIEW OF LITERATURE Author Year Findings Ganguli and Reddy 2014 SVM is able to learn from a much smaller data set and is capable of handling a large number of variables. SVM may overcome certain limitations of ANN such as trapping at local minima and overfitting, to some extent in predicting droughts Yang et al 2015 Among the ML techniques, ANN and SVM can be regarded as the most widely used techniques for developing drought prediction models
  • 4. Study area and datasets • Pakistan located in South Asia covers an area of 796,095 km2. • Pakistan experiences two monsoon precipitation seasons; IM (July–Sep), WD (Dec–Mar). Fig. 1. IM refers to the regions which receive precipitation mainly during the Indian monsoon and WD refers to the regions which receive precipitation mainly during the western disturbance.
  • 5. • It experiences four seasons (based on temperature) cool and dry winter (Dec–Feb), hot and dry spring (Mar–May), hot and humid summer (Jun–Aug) and dry autumn (Sep–Nov) • During these two precipitation seasons, two important cropping seasons called Kharif (May-Oct) observed during the Indian monsoon period • The Rabi (Nov-Apr) is experienced during the western disturbances • These cropping seasons play an important role in agriculture and the economy of Pakistan
  • 6. Gridded precipitation and temperature datasets • This study uses the Princeton global forcing (PGF) gridded temperature and precipitation datasets for estimating the SPEI (NCEP/NCAR datasets) • The PGF gridded temperature and precipitation datasets have a resolution of 0.25 ×0.25 in the latitudinal and longitudinal directions and hence the data are available at 1437 grid points over Pakistan. • These data were used in the calculation of SPEI at each of the 1437 grid points distributed over Pakistan, Among which 531 grid points lie in the region of WD and 906 grid points lie in the IM dominated region
  • 7. Method 1–6-month SPEIs were calculated for each of the grid points Based on 1–6-month SPEIs droughts were categorized Different large-scale atmospheric variables (probable predictors) that are influential on droughts were identified based on past literature Principle Component Analysis (PCA) was used to generate the Principal Components (PCs) from the data The difference between the PCs for February and April(Kharif) and difference between the PCs for October and August(Rabi) were computed. The PCs selected using SVM-RFE they were used in the development of drought prediction models the correlations between the PCs and 1–6-month SPEI were separately calculated ML-algorithms were used to develop drought prediction
  • 8. PCA Why? • Overfitting (Due to too much features and attribute we gave during training phase so model is confused and problem of overfitting occur) • PCA is Linear projection method to reduce number of parameters(Attributes and features). 1. Standardize the data 2. Calculate the covariance matrix 3. Find the eigenvalues and eigenvectors of covariance matrix 4. Plot the eigenvectors/principal components over scaled data
  • 9. SVM • SVM is a widely used ML algorithm that can be used in developing classification and regression models. • However, Sachindra et al. (2018) showed that the polynomial kernel performs better with SVM. Hence, the polynomial kernel was used in this study. • The polynomial kernel is defined as . 𝐾 (𝑥 𝑖 , 𝑥 𝑗 )= (𝑥 𝑖 . 𝑥 𝑗 + 𝑐 ) 𝑑 Where, 𝑥 𝑖 and 𝑥 𝑗 are the predictor and predictand data, d is the degree of the polynomial, and c is a constant that allows a trade-off between the influence of the higher and lower order terms.
  • 10. ANN • .To reduce the chances of overfitting, a technique known as Bayesian regularization was developed ( Burden and Winkler, 2008 ; MacKay, 1992 ; Ticknor, 2013 ). 𝐹 = 𝛽E D + 𝛼E w • In the Bayesian network, the weights are considered random variables and thus their density function is expressed following the Bayes’ rules 𝑃 ( 𝑤 |𝐷, 𝛼, 𝛽, M) = 𝑃 (D |w , 𝛽, M) P(w |𝛼, M) 𝑃 (D |𝛼, 𝛽, M) Where w is the vector of network weights, D represents the predictor and the predictand data vectors (x, y), and M is the neural network model being used, 𝛼 and 𝛽 are Regularization
  • 11. K-Nearest Neighbour • The KNN for classification and regression is the simplest non-parametric ML techniques • KNN algorithm finds the k Nearest Neighbours from calibration data using a distance measure such as the Euclidean distance 𝑑 𝑗𝑜 = 𝑖=1 ݊ (𝑥 𝑖𝑗 − 𝑥 𝑖𝑜 )2 𝑡 = 1 , 2 , 3 , …., ݊ • For calculating KNN simulated values ( Zr ) Zk refers to the neighbouring data and f k ( dj ) is the kernel function 𝑍 𝑟 = ݇ =1 𝐾 ݂݇ (𝑑𝑗 ×𝑍݇ ).
  • 12. Performance assessment 1) The coefficient of determination (R2 ) varies between 0 (no agreement) and 1 (perfect agreement) 2) The normalized root mean squared error (NRMSE) varies between 0 (perfect agreement) and +∞(no agreement) 3) Percentage of bias (Pbias) varies between - ∞(underestimation) and +∞(over- estimation) and a value of 0 refers to a perfect agreement 4) lThe modified index of agreement ( md ) varies between 0 (no agreement) and 1 (perfect agreement) N is the number of testing samples. Oi are the ith observation, Si are the i th simulated value, sd is the standard deviation of the observations.
  • 13. Results and Discussion Fig. 2. Climate domain used in the present study, 861 NCEP/NCAR grid points spread across the climate domain are depicted in grey circles.
  • 14. Fig. 3. The spatial pattern of the correlation coefficient of the RH, atm temperature and wind component (925hPa) during the Rabi and Kharif periods. The correlation is calculated with 6-month SPEI in April for Rabi and in October for Kharif. The dots represent the significant correlation at 95% confidence level with the SPEI.
  • 15. Fig. 4. Boxplots showing correlations of the selected atmospheric variables at four pressure levels i.e. 925, 850, 700 and 500 hPa (except SLP) with the SPEI of Pakistan Red boxplots (Rabi season) Blue (Kharif season)
  • 16. Fig. 5. Performance of different machine learning-based models in predicting moderate droughts in the Rabi season.
  • 17. Fig. 6. Performance of different machine learning-based models in predicting moderate droughts in the Kharif season.
  • 18. Fig.7. Boxplots of the spatial correlation and NRMSE for the Rabi and Kharif seasons for the month of April shown in red and blue colours respectively.
  • 19. Conclusions • In this study, it was seen that KNN-based drought models display limited performance comparison to other drought models • SVM-based models were able to better capture the temporal and spatial characteristics of droughts • It was found that in the Rabi season SPEI is positively correlated with RH over the Mediterranean Sea and the region north of the Caspian Sea, whereas in the Kharif season SPEI is positively correlated with the humid region over the south-eastern part of the Bay of Bengal and Caspian Seas.
  • 20. • Rabi season SPEI also displayed statistically significant positive correlations with atmospheric temperature over the region north of the Mediterranean Sea. • Also, wind speeds over the Indian Ocean and the Arabian Sea during the Kharif season were statistically significantly correlated with SPEI . • This indicated that RH, temperature and wind speed are indicators of droughts over Pakistan.
  • 21. Limitations of this study • The present study used only one reanalysis dataset (NCEP/NCAR) to provide inputs to the drought prediction models. • In order to study the uncertainties which originate from the inputs, different reanalysis datasets should be used in developing drought prediction models. • However, this was not performed as the other major reanalysis datasets either did not have data for some of the specific atmospheric variables used in this study or they did not have data corresponding to the period 1948– 2016.
  • 22. Future scope • Sachindra and Kanae (2019) proved that Parallel Multi-Population Genetic Programming (PMPGP)-based statistical downscaling models show better generalization skills and higher resistance to redundant information in inputs compared to those based on traditional GP. • Thus, other ML algorithms such as Extreme Learning Machine (ELM), Genetic Programming (GP), PMPGP, and Random Forest (RF) should also be tested in future for developing drought prediction models over Pakistan.
  • 23. References • Najeebullah Khan, D.A. Sachindra , Shamsuddin Shahid , Kamal Ahmed a , Mohammed Sanusi Shiru , Nadeem Nawaz “Prediction of droughts over Pakistan using machine learning algorithms ”, Advances in Water Resources(2020) Volume 139, May 2020, 103562 • Ganguli, P. , Reddy, M.J. , 2014. Ensemble prediction of regional droughts using climate inputs and the SVM–copula approach. Hydrol. Process. 28, 4989–5009