Abstract:As the role played by statistical and computational sciences in climate and environmental modelling and prediction becomes more important, Machine Learning researchers are becoming more aware of the relevance of their work to help tackle the climate crisis. Indeed, being universal nonlinear function approximation tools, Machine Learning algorithms are efficient in analysing and modelling spatially and temporally variable environmental data. While Deep Learning models have proved to be able to capture spatial, temporal, and spatio-temporal dependencies through their automatic feature representation learning, the problem of the interpolation of continuous spatio-temporal fields measured on a set of irregular points in space is still under-investigated. To fill this gap, we introduce here a framework for spatio-temporal prediction of climate and environmental data using deep learning. Specifically, we show how spatio-temporal processes can be decomposed in terms of a sum of products of temporally referenced basis functions, and of stochastic spatial coefficients which can be spatially modelled and mapped on a regular grid, allowing the reconstruction of the complete spatio-temporal signal. Applications on two case studies based on simulated and real-world data will show the effectiveness of the proposed framework in modelling coherent spatio-temporal fields.

What problem does this paper attempt to address?

The paper attempts to address the problem of how to use deep learning methods to interpolate irregularly distributed ground measurement data for reconstructing continuous spatiotemporal fields when performing spatiotemporal predictions in environmental data. Specifically, the paper proposes a deep learning-based framework for reconstructing spatiotemporal fields from irregularly distributed time series data and interpolating them on a regular grid. The background of this problem is that traditional deep learning methods are mainly applied to regular grid data (such as satellite data), while the interpolation problem for irregularly distributed ground measurement data has not been fully studied. The main contributions of the paper include: 1. **Proposing a new framework**: This framework decomposes the spatiotemporal process, representing it as a sum of products of temporal basis functions and spatial coefficients, and then uses deep neural networks to model these spatial coefficients. 2. **Solving the interpolation problem for irregularly distributed data**: Traditional machine learning methods usually can only model at the locations of measurement stations and cannot interpolate at arbitrary locations. The proposed framework can interpolate at any location, thereby better utilizing ground measurement data. 3. **Validating the effectiveness of the method**: Through two case studies (a simulated dataset and an actual temperature monitoring network dataset), the effectiveness of the framework in reconstructing spatiotemporal fields is demonstrated. In summary, this paper aims to fill the gap in existing methods when dealing with irregularly distributed ground measurement data, providing a new solution for spatiotemporal predictions of climate and environmental data.

A Novel Framework for Spatio-Temporal Prediction of Environmental Data Using Deep Learning

STONet: A Neural-Operator-Driven Spatio-temporal Network

Deep learning for spatiotemporal forecasting in Earth system science: a review

A spatio-temporal LSTM model to forecast across multiple temporal and spatial scales

Deep learning and process understanding for data-driven Earth system science

Predicting extreme events from data using deep machine learning: when and where

A Deep Learning Approach for Predicting Spatiotemporal Dynamics From Sparsely Observed Data

Spatio temporal hydrological extreme forecasting framework using LSTM deep learning model

Graph-Based Deep Modeling and Real Time Forecasting of Sparse Spatio-Temporal Data.

Micro-climate Prediction - Multi Scale Encoder-decoder based Deep Learning Framework

Building Autocorrelation-Aware Representations for Fine-Scale Spatiotemporal Prediction

Deep learning and a changing economy in weather and climate prediction

DNN-based Prediction Model for Spatio-Temporal Data

Deep Learning Based Prediction Of Weather Using Hybrid_stacked Bi-Long Short Term Memory

Spatially-distributed Deep Learning for rainfall-runoff modelling and system understanding

Advances and prospects of deep learning for medium-range extreme weather forecasting

DEFM: Delay E mbedding based Forecast Machine for Time Series Forecasting by Spatiotemporal Information Transformation

Unravelled multilevel transformation networks for predicting sparsely observed spatio-temporal dynamics

A Spatiotemporal Prediction Framework for Air Pollution Based on Deep RNN

Deep learning framework with Bayesian data imputation for modelling and forecasting groundwater levels

DeepOcean: A General Deep Learning Framework for Spatio-Temporal Ocean Sensing Data Prediction.