Stochastic Downscaling to Chaotic Weather Regimes using Spatially Conditioned Gaussian Random Fields with Adaptive Covariance

Downscaling aims to link the behaviour of the atmosphere at fine scales to properties measurable at coarser scales, and has the potential to provide high resolution information at a lower computational and storage cost than numerical simulation alone. This is especially appealing for targeting convective scales, which are at the edge of what is possible to simulate operationally. Since convective scale weather has a high degree of independence from larger scales, a generative approach is essential. We here propose a statistical method for downscaling moist variables to convective scales using conditional Gaussian random fields, with an application to wet bulb potential temperature (WBPT) data over the UK. Our model uses an adaptive covariance estimation to capture the variable spatial properties at convective scales. We further propose a method for the validation, which has historically been a challenge for generative models.

[1]  A. Raftery,et al.  Probabilistic forecasts, calibration and sharpness , 2007 .

[2]  D. Caya,et al.  Spatial Disaggregation of Mean Areal Rainfall Using Gibbs Sampling , 2012 .

[3]  M. Schirrmann,et al.  Area-to-Point Kriging of Soil Phosphorus Composite Samples , 2012 .

[4]  Subimal Ghosh,et al.  SVM-PGSL coupled approach for statistical downscaling to predict rainfall from GCM output , 2010 .

[5]  P. Whetton,et al.  Guidelines for Use of Climate Scenarios Developed from Statistical Downscaling Methods , 2004 .

[6]  B. Brown,et al.  Object-Based Verification of Precipitation Forecasts. Part I: Methodology and Application to Mesoscale Rain Areas , 2006 .

[7]  L. Mark Berliner,et al.  Spatiotemporal Hierarchical Bayesian Modeling Tropical Ocean Surface Winds , 2001 .

[8]  M. Xue,et al.  Using new neighborhood-based intensity-scale verification metrics to evaluate WRF precipitation forecasts at 4 and 12 km grid spacings , 2020 .

[9]  D. Cox,et al.  An Analysis of Transformations , 1964 .

[10]  Richard Swinbank,et al.  MOGREPS-UK Convection-Permitting Ensemble Products for Surface Water Flood Forecasting: Rationale and First Results , 2016 .

[11]  Leonhard Held,et al.  Gaussian Markov Random Fields: Theory and Applications , 2005 .

[12]  Y. LindaJ. Combining Incompatible Spatial Data , 2003 .

[13]  E. Foufoula‐Georgiou,et al.  Model for multiscale disaggregation of spatial rainfall based on coupling meteorological and scaling descriptions , 1996 .

[14]  Xiang Yu,et al.  Ensemble spatiotemporal forecasting of solar irradiation using variational Bayesian convolutional gate recurrent unit network , 2019, Applied Energy.

[15]  Chris A. Glasbey,et al.  A latent Gaussian Markov random‐field model for spatiotemporal rainfall disaggregation , 2003 .

[16]  Yann LeCun,et al.  Deep multi-scale video prediction beyond mean square error , 2015, ICLR.

[17]  R. Katz,et al.  Daily spatiotemporal precipitation simulation using latent and transformed Gaussian processes , 2012 .

[18]  C. Villani Optimal Transport: Old and New , 2008 .

[19]  Marco Cuturi,et al.  On Wasserstein Two-Sample Testing and Related Families of Nonparametric Tests , 2015, Entropy.

[20]  P. Rowntree,et al.  A Mass Flux Convection Scheme with Representation of Cloud Ensemble Characteristics and Stability-Dependent Closure , 1990 .

[21]  Alain Trouvé,et al.  Interpolating between Optimal Transport and MMD using Sinkhorn Divergences , 2018, AISTATS.

[22]  Thomas Vandal,et al.  Intercomparison of machine learning methods for statistical downscaling: the case of daily and extreme precipitation , 2017, Theoretical and Applied Climatology.

[23]  M. Schlather,et al.  A Matern based multivariate Gaussian random process for a consistent model of the horizontal wind components and related variables , 2017, 1707.01287.

[24]  Christopher J. Paciorek,et al.  Nonstationary Gaussian Processes for Regression and Spatial Modelling , 2003 .

[25]  D. Stephenson,et al.  A new intensity‐scale approach for the verification of spatial precipitation forecasts , 2004 .

[26]  Nicola Rebora,et al.  RainFARM: Rainfall Downscaling by a Filtered Autoregressive Model , 2006 .

[27]  Yan Jin,et al.  Principles and methods of scaling geospatial Earth science data , 2019, Earth-Science Reviews.

[28]  H. V. D. Dool,et al.  Searching for analogues, how long must we wait? , 1994 .

[29]  J. Wakefield,et al.  Modeling Spatial Variation in Disease Risk , 2002 .

[30]  N. Roberts,et al.  Scale-Selective Verification of Rainfall Accumulations from High-Resolution Forecasts of Convective Events , 2008 .

[31]  Martyn P. Clark,et al.  HEPEX: The Hydrological Ensemble Prediction Experiment , 2007 .

[32]  G. Strang Wavelet transforms versus Fourier transforms , 1993, math/9304214.

[33]  Paolo Burlando,et al.  Stochastic downscaling of climate model precipitation outputs in orographically complex regions: 2. Downscaling methodology , 2014 .

[34]  D. Nychka,et al.  A Multiresolution Gaussian Process Model for the Analysis of Large Spatial Datasets , 2015 .

[35]  Sangram Ganguly,et al.  DeepSD: Generating High Resolution Climate Change Projections through Single Image Super-Resolution , 2017, KDD.

[36]  Zbynek Sokol,et al.  A radar-based verification of precipitation forecast for local convective storms , 2007 .

[37]  Chuong B. Do More on Multivariate Gaussians , 2008 .

[38]  Luca Delle Monache,et al.  Probabilistic Weather Prediction with an Analog Ensemble , 2013 .

[39]  U. Germann,et al.  Pysteps: an open-source Python library for probabilistic precipitation nowcasting (v1.0) , 2019, Geoscientific Model Development.

[40]  Léon Bottou,et al.  Wasserstein Generative Adversarial Networks , 2017, ICML.

[41]  Edward C. Waymire,et al.  A statistical analysis of mesoscale rainfall as a random cascade , 1993 .

[42]  Roger Woodard,et al.  Interpolation of Spatial Data: Some Theory for Kriging , 1999, Technometrics.

[43]  Mladen Kezunovic,et al.  Spatial-temporal solar power forecast through use of Gaussian Conditional Random Fields , 2016, 2016 IEEE Power and Energy Society General Meeting (PESGM).

[44]  Mario Chica-Olmo,et al.  Downscaling cokriging for image sharpening , 2006 .

[45]  P. Kyriakidis A Geostatistical Framework for Area-to-Point Spatial Interpolation , 2004 .

[46]  Maogui Hu,et al.  atakrig: An R package for multivariate area-to-area and area-to-point kriging predictions , 2020, Comput. Geosci..

[47]  A. Arakawa,et al.  Interaction of a Cumulus Cloud Ensemble with the Large-Scale Environment, Part I , 1974 .

[48]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[49]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[50]  A new discrete multiplicative random cascade model for downscaling intermittent rainfall fields , 2020 .

[51]  S. Pryor,et al.  Downscaling temperature and precipitation: a comparison of regression‐based methods and artificial neural networks , 2001 .

[52]  Elizabeth E. Ebert,et al.  Fuzzy verification of high‐resolution gridded forecasts: a review and proposed framework , 2008 .

[53]  Taha B. M. J. Ouarda,et al.  Automated regression-based statistical downscaling tool , 2008, Environ. Model. Softw..

[54]  Marc G. Bellemare,et al.  The Cramer Distance as a Solution to Biased Wasserstein Gradients , 2017, ArXiv.

[55]  Makarand Tapaswi,et al.  A Closed-form Gradient for the 1 D Earth Mover ’ s Distance for Spectral Deep Learning on Biological Data , 2016 .

[56]  S. L. Eerm,et al.  Fractal properties of rain, and a fractal model , 1985 .

[57]  H. Storch,et al.  The Analog Method as a Simple Statistical Downscaling Technique: Comparison with More Complicated Methods , 1999 .

[58]  C. Jakob,et al.  Stochastic Space‐Time Downscaling of Rainfall Using Event‐Based Multiplicative Cascade Simulations , 2019, Journal of Geophysical Research: Atmospheres.

[59]  R. Plant,et al.  Characterisation of convective regimes over the British Isles , 2016 .

[60]  Xueyou Li,et al.  Using Conditioned Random Field to Characterize the Variability of Geologic Profiles , 2016 .

[61]  J. Zico Kolter,et al.  Large-scale probabilistic forecasting in energy systems using sparse Gaussian conditional random fields , 2013, 52nd IEEE Conference on Decision and Control.

[62]  Richard Swinbank,et al.  The Met Office convective‐scale ensemble, MOGREPS‐UK , 2017 .

[63]  M. Martínez-Beneito A general modelling framework for multivariate disease mapping , 2013 .

[64]  Eric Gilleland,et al.  Intercomparison of Spatial Forecast Verification Methods , 2009 .

[65]  Arun Kumar,et al.  Long‐range experimental hydrologic forecasting for the eastern United States , 2002 .