SASCOF 2022: Seasonal Forecasts of Summer Monsoon Rainfall with Extreme Learning Machine
By Nachiketa Acharya & Kyle Hall
Here we present both deterministic and probabilistic seasonal forecasts of Summer Monsoon Rainfall (SMR) over the South Asia region. This rainfall occurs during the JJAS season, and is characterized by high interannual variability, which presents a significant forecasting challenge.
These forecasts are produced using Extreme Learning Machine (ELM) based Multi-Model Ensembles (Acharya et al,2014). ELM is a neural network-based non-linear regression method, which serves as a fast alternative to the traditional Single Layer Feed-Forward neural network (Huang et al,2004). Since ELM is designed to produce deterministic forecasts, we have used a modified version known as Probabilistic Output ELM (POELM) to produce probabilistic forecasts (Wong et al.,2020). POELM replaces the linear objective function of ELM with a sigmoid objective function, and uses sigmoid additive hidden neurons. Here, we have also altered the hidden neurons of POELM to use different types of activation functions, and adopted a postprocessing rule using normalization and softmax to ensure that the tercile probabilities are reasonable. (Pending publication)
Summer Monsoon Rainfall 2022 Forecast:
Experimental Design:
Predictors: 4 North American Multi-Model Ensemble (NMME) models - CCSM4, CFSv2, NASA-GEOS, and CANSIPS-IC3
Predictand: SASCOF observed JJAS precipitation
Training Period: April Initializations from 1982-2019
Hyper-Parameters:
- 5 hidden layer neurons using ReLu activation
- MinMax scaling to (-1, 1) of predictors
- Simple anomaly scaling of predictand
- One-Hot Encoding of predictand for POELM
- Cross Validation Window: 3 years (Leave-3yrs-Out Crossvalidation)
- 3x3 Gaussian Kernel Smoothing of output
Leave-3-Out Cross-Validated Skill Metrics
References:
Acharya N, Srivastava N.A., Panigrahi B.K. and Mohanty U.C. (2014): Development of an artificial neural network based multi- model ensemble to estimate the northeast monsoon rainfall over south peninsular India: an application of extreme learning machine. Climate Dynamics. 43(5):1303-1310.
Guang-Bin Huang, Qin-Yu Zhu, & Chee-Kheong Siew. (2004). Extreme learning machine: A new learning scheme of feedforward neural networks. 2, 985–990 vol.2. https://doi.org/ 10.1109/IJCNN.2004.1380068.
Wong SY, Yap KS, Li X (2020) A new probabilistic output constrained optimization extreme learning machine. IEEE Access 8:28,934– 28,946.