Meteorological drought forecasting for ungauged areas based on machine learning: Using long-range climate forecast and remote sensing data

Rhee, Jinyoung; Im, Jungho

doi:10.1016/j.agrformet.2017.02.011

Cited by 185 publications

(109 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In many cases, the complicated interaction in the hydroclimatic system cannot be characterized by linear models. The Artificial Intelligence (AI) (or machine learning, soft computing) models, including Artificial Neural Network (ANN), Fuzzy Logic (FL), Support Vector Regression (SVR) or Support Vector Machine (SVM), Genetic Algorithm (GA) or Genetic Programming (GP), and wavelet transformation, can be used to model complex interactions of hydroclimatic variables for a variety of applications (Bourdin et al, 2012;Fahimi et al, 2016;Nourani et al, 2014;Rhee & Im, 2017;Wang, Chau, et al, 2009;Yaseen et al, 2015). Several AI models, including the ANN (Mishra & Desai, 2006;Mishra et al, 2007;Morid et al, 2007), SVM (Ganguli & Reddy, 2014), and wavelet transformation (Maity et al, 2016;Özger et al, 2011), have been used to model complicated and nonlinear interactions between drought indicators and influencing factors for drought prediction.…”

Section: Artificial Intelligence Modelmentioning

confidence: 99%

Seasonal Drought Prediction: Advances, Challenges, and Future Prospects

Hao

Singh

Xia

2018

Reviews of Geophysics

390

232

View full text Add to dashboard Cite

Drought prediction is of critical importance to early warning for drought managements. This review provides a synthesis of drought prediction based on statistical, dynamical, and hybrid methods. Statistical drought prediction is achieved by modeling the relationship between drought indices of interest and a suite of potential predictors, including large‐scale climate indices, local climate variables, and land initial conditions. Dynamical meteorological drought prediction relies on seasonal climate forecast from general circulation models (GCMs), which can be employed to drive hydrological models for agricultural and hydrological drought prediction with the predictability determined by both climate forcings and initial conditions. Challenges still exist in drought prediction at long lead time and under a changing environment resulting from natural and anthropogenic factors. Future research prospects to improve drought prediction include, but are not limited to, high‐quality data assimilation, improved model development with key processes related to drought occurrence, optimal ensemble forecast to select or weight ensembles, and hybrid drought prediction to merge statistical and dynamical forecasts.

show abstract

Section: Artificial Intelligence Modelmentioning

confidence: 99%

Seasonal Drought Prediction: Advances, Challenges, and Future Prospects

Hao

Singh

Xia

2018

Reviews of Geophysics

390

232

View full text Add to dashboard Cite

show abstract

“…This is possibly because a smaller number of training samples were available for the Himawari-8 models (Table 3), which often results in overfitting [62,63]. In particular, the ranges of the input values of the icing samples for the Himawari-8-based models were generally smaller than those for COMS-based models.…”

Section: Model Performancementioning

confidence: 97%

“…However, only a small number of icing PIREPs are available in East Asia, which is a major limitation of this research. Such a small sample size often results in overfitting [62,63]. Since satellite-derived products typically provide information on the characteristics of cloud tops, it should be noted that PIREP-based icing information collected far below the cloud tops is not always closely related to satellite-derived cloud top products.…”

Section: Novelty and Limitationsmentioning

confidence: 99%

“…The mean decrease accuracy (MDA) values calculated when a variable is randomly permuted were used to identify significant variables in the RF model, while the absolute normalized coefficient values were used in the MLL model [62,63]. It should be noted that these metrics suggest the relative importance of variables in estimating the target variable, not the absolute indicator of variable significance.…”

Section: Variable Importancementioning

confidence: 99%

See 1 more Smart Citation

Icing Detection over East Asia from Geostationary Satellite Data Using Machine Learning Approaches

Sim

Park

et al. 2018

Remote Sensing

Self Cite

View full text Add to dashboard Cite

Even though deicing or airframe coating technologies continue to develop, aircraft icing is still one of the critical threats to aviation. While the detection of potential icing clouds has been conducted using geostationary satellite data in the US and Europe, there is not yet a robust model that detects potential icing areas in East Asia. In this study, we proposed machine-learning-based icing detection models using data from two geostationary satellites-the Communication, Ocean, and Meteorological Satellite (COMS) Meteorological Imager (MI) and the Himawari-8 Advanced Himawari Imager (AHI)-over Northeast Asia. Two machine learning techniques-random forest (RF) and multinomial log-linear (MLL) models-were evaluated with quality-controlled pilot reports (PIREPs) as the reference data. The machine-learning-based models were compared to the existing models through five-fold cross-validation. The RF model for COMS MI produced the best performance, resulting in a mean probability of detection (POD) of 81.8%, a mean overall accuracy (OA) of 82.1%, and mean true skill statistics (TSS) of 64.0%. One of the existing models, flight icing threat (FIT), produced relatively poor performance, providing a mean POD of 36.4%, a mean OA of 61.0, and a mean TSS of 9.7%. The Himawari-8 based models also produced performance comparable to the COMS models. However, it should be noted that very limited PIREP reference data were available especially for the Himawari-8 models, which requires further evaluation in the future with more reference data. The spatio-temporal patterns of the icing areas detected using the developed models were also visually examined using time-series satellite data.

show abstract

“…Breiman [46] provides more details about RF. RF has been widely used for various classification and regression tasks [51][52][53][54][55][56][57][58][59][60][61][62] and is known to better overcome an overfitting problem than simple decision trees such as CART. RF requires less setting of parameters and is faster than SVM and other ensemble classifiers [57].…”

Section: Random Forestmentioning

confidence: 99%

Estimation of Fugacity of Carbon Dioxide in the East Sea Using In Situ Measurements and Geostationary Ocean Color Imager Satellite Data

Jang

Park

et al. 2017

Remote Sensing

Self Cite

View full text Add to dashboard Cite

Abstract:The ocean is closely related to global warming and on-going climate change by regulating amounts of carbon dioxide through its interaction with the atmosphere. The monitoring of ocean carbon dioxide is important for a better understanding of the role of the ocean as a carbon sink, and regional and global carbon cycles. This study estimated the fugacity of carbon dioxide (ƒCO 2 ) over the East Sea located between Korea and Japan. In situ measurements, satellite data and products from the Geostationary Ocean Color Imager (GOCI) and the Hybrid Coordinate Ocean Model (HYCOM) reanalysis data were used through stepwise multi-variate nonlinear regression (MNR) and two machine learning approaches (i.e., support vector regression (SVR) and random forest (RF)). We used five ocean parameters-colored dissolved organic matter (CDOM; <0.3 m −1 ), chlorophyll-a concentration (Chl-a; <21 mg/m 3 ), mixed layer depth (MLD; <160 m), sea surface salinity (SSS; 32-35), and sea surface temperature (SST; 8-28 • C)-and four band reflectance (Rrs) data (400 nm-565 nm) and their ratios as input parameters to estimate surface seawater ƒCO 2 (270-430 µatm). Results show that RF generally performed better than stepwise MNR and SVR. The root mean square error (RMSE) of validation results by RF was 5.49 µatm (1.7%), while those of stepwise MNR and SVR were 10.59 µatm (3.2%) and 6.82 µatm (2.1%), respectively. Ocean parameters (i.e., sea surface salinity (SSS), sea surface temperature (SST), and mixed layer depth (MLD)) appeared to contribute more than the individual bands or band ratios from the satellite data. Spatial and seasonal distributions of monthly ƒCO 2 produced from the RF model and sea-air CO 2 flux were also examined.

show abstract

Meteorological drought forecasting for ungauged areas based on machine learning: Using long-range climate forecast and remote sensing data

Cited by 185 publications

References 50 publications

Seasonal Drought Prediction: Advances, Challenges, and Future Prospects

Seasonal Drought Prediction: Advances, Challenges, and Future Prospects

Icing Detection over East Asia from Geostationary Satellite Data Using Machine Learning Approaches

Estimation of Fugacity of Carbon Dioxide in the East Sea Using In Situ Measurements and Geostationary Ocean Color Imager Satellite Data

Contact Info

Product

Resources

About