Deep Integro-Difference Equation Models for Spatio-Temporal Forecasting

by   Andrew Zammit Mangion, et al.

Integro-difference equation (IDE) models describe the conditional dependence between the spatial process at a future time point and the process at the present time point through an integral operator. Nonlinearity or temporal dependence in the dynamics is often captured by allowing the operator parameters to vary temporally, or by re-fitting a model with a temporally-invariant linear operator at each time point in a sliding window. Both procedures tend to be excellent for prediction purposes over small time horizons, but are generally time-consuming and, crucially, do not provide a global prior model for the temporally-varying dynamics that is realistic. Here, we tackle these two issues by using a deep convolution neural network (CNN) in a hierarchical statistical IDE framework, where the CNN is designed to extract process dynamics from the process' most recent behaviour. Once the CNN is fitted, probabilistic forecasting can be done extremely quickly online using an ensemble Kalman filter with no requirement for repeated parameter estimation. We conduct an experiment where we train the model using 13 years of daily sea-surface temperature data in the North Atlantic Ocean. Forecasts are seen to be accurate and calibrated. A key advantage of our approach is that the CNN provides a global prior model for the dynamics that is realistic, interpretable, and computationally efficient. We show the versatility of the approach by successfully producing 10-minute nowcasts of weather radar reflectivities in Sydney using the same model that was trained on daily sea-surface temperature data in the North Atlantic Ocean.


page 20

page 25


An efficient estimation of spatio-temporally distributed parameters in dynamic models by an ensemble Kalman filter

The accuracy of Earth system models is compromised by unknown and/or unr...

A Deep Convolutional Neural Network Model for improving WRF Forecasts

Advancements in numerical weather prediction models have accelerated, fo...

Learning the Dynamics of Physical Systems from Sparse Observations with Finite Element Networks

We propose a new method for spatio-temporal forecasting on arbitrarily d...

Daily Data Assimilation of a Hydrologic Model Using the Ensemble Kalman Filter

Accurate runoff forecasting is crucial for reservoir operators as it all...

D-Vine GAM Copula based Quantile Regression with Application to Ensemble Postprocessing

Temporal, spatial or spatio-temporal probabilistic models are frequently...

Learning Insulin-Glucose Dynamics in the Wild

We develop a new model of insulin-glucose dynamics for forecasting blood...

Please sign up or login with your details

Forgot password? Click here to reset