Asee peer logo

Forecasting Drought Indices Using Machine Learning Algorithm

Download Paper |

Conference

2020 ASEE Virtual Annual Conference Content Access

Location

Virtual On line

Publication Date

June 22, 2020

Start Date

June 22, 2020

End Date

June 26, 2021

Conference Session

Instrumentation Division Technical Session 2

Tagged Division

Instrumentation

Tagged Topic

Diversity

Page Count

11

DOI

10.18260/1-2--34680

Permanent URL

https://peer.asee.org/34680

Download Count

963

Request a correction

Paper Authors

biography

Jay Lee P.E. California Baptist University

visit author page

Dr. Lee’s research interest is in information technology and strategic decision-making practices in various engineering management fields. His current research topics include spatial analysis utilizing a commercial Geographic Information System (GIS) applications and machine learning-based forecasting in engineering practices.

visit author page

biography

Yeonsang Hwang P.E. Arkansas State University

visit author page

Associate Professor of Civil Engineering

visit author page

biography

Tae-Hoon Kim Purdue University Northwest Orcid 16x16 orcid.org/0000-0001-8595-7925

visit author page

Dr. Tae-Hoon Kim, Ph.D in Information Science, associate professor of Computer Information Technology and Graphics. His teaching areas are computer networking, network security, network design, parallel computing, and data science. His research interests are reliable wireless sensor and ad hoc network, network anomaly detection, cyber-physical system, and applied data science.

visit author page

Download Paper |

Abstract

According to the existing studies, the historical climate record and seasonal temperature and precipitation records offer useful input for making short-term drought predictions. In the last few decades, numerous studies have been conducted to explore these data in a way to predict upcoming drought events. Despite the efforts, few studies have succeeded in quantifying uncertainties in the process of predicting drought index values due mainly to technical challenges and implications in computation. This paper proposes a new approach utilizing an artificial intelligence model for forecasting drought indices. This study uses a regression analysis model in machine learning, Lasso, which is normalized to improve the prediction accuracy. Lasso model will be implemented in Python using scikit-learn, and 10-fold cross-validation will be used to ensure the prediction accuracy. The proposed model uses the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center (CPC) seasonal data to compute the Palmer Drought Severity Index (PDSI). The accuracy of the model is validated using the historical records of drought indices and available seasonal temperature and precipitation data provided by the NOAA CPC. The results of the forecasts produced by this model will be compared with the observed drought indices and validated. The mean error rate and root mean square error (RMSE) methods are used to measure the accuracy of the forecast at stations for validation. The validated model can be used in classroom and laboratory settings for general engineering studies.

Lee, J., & Hwang, Y., & Kim, T. (2020, June), Forecasting Drought Indices Using Machine Learning Algorithm Paper presented at 2020 ASEE Virtual Annual Conference Content Access, Virtual On line . 10.18260/1-2--34680

ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2020 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015