Statistical downscaling of precipitation using long short-term memory recurrent neural networks

Misra, Saptarshi; Sarkar, Sudeshna; Mitra, Pabitra

doi:10.1007/s00704-017-2307-2

Cited by 68 publications

(43 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, long-term memory implies that the present state of the system influences the future states. Recent advancements in deep learning literature result in several SD architectures such as Convolution Neural Networks (CNNs) [22], Residual Dense Block (RDB) [23] and Long Short Term Memory (LSTMs) [24] capturing spatial and temporal dependencies, respectively.…”

Section: Spatio-temporal Teleconnectionsmentioning

confidence: 99%

Augmented Convolutional LSTMs for Generation of High-Resolution Climate Change Projections

2021

View full text Add to dashboard Cite

Projection of changes in extreme indices of climate variables such as temperature and precipitation are critical to assess the potential impacts of climate change on human-made and natural systems, including critical infrastructures and ecosystems. While impact assessment and adaptation planning rely on high-resolution projections (typically in the order of a few kilometers), state-of-the-art Earth System Models (ESMs) are available at spatial resolutions of few hundreds of kilometers. Current solutions to obtain highresolution projections of ESMs include downscaling approaches that consider the information at a coarsescale to make predictions at local scales. Complex and non-linear interdependence among local climate variables (e.g., temperature and precipitation) and large-scale predictors (e.g., pressure fields) motivate the use of neural network-based super-resolution architectures. In this work, we present auxiliary variables informed spatio-temporal neural architecture for statistical downscaling. The current study performs daily downscaling of precipitation variable from an ESM output at 1.15 degrees (115 km) to ¼ degrees (25 km) over the one of the most climatically diversified countries, India. We showcase significant improvement gain against two popular state-of-the-art baselines with a better ability to predict statistics of extreme events. To facilitate reproducible research, we make available all the codes, processed datasets, and trained models in the public domain.

show abstract

Section: Spatio-temporal Teleconnectionsmentioning

confidence: 99%

Augmented Convolutional LSTMs for Generation of High-Resolution Climate Change Projections

2021

View full text Add to dashboard Cite

show abstract

“…A limited amount of machine learning studies has been reported in terms of downscaling of air temperatures Baño-Medina et al, 2019; Sachindra and Kanae, 2019 , whereas considerable success has been achieved in downscaling of precipitation e.g., Wilby et al, 1998;Vandal et al, 2017;Misra et al, 2018;Baño-Medina et al, 2019;Pan et al, 2019 . Gaps between model formulas and actual events of rainfall processes may be greater than those in temperature, and hence, there may be still room for the contribution of machine learning.…”

Section: Use Of Machine Learning Techniques In Meteorologymentioning

confidence: 99%

“…Another important moiety to be addressed is spatial estimation, namely, downscaling. Several studies have attempted to apply these techniques to meteorological downscaling and demonstrated their effectiveness in terms of air temperatures Baño-Medina et al, 2019;Sachindra andKanae, 2019 , precipitation e.g., Wilby et al, 1998;Vandal et al, 2017;Misra et al, 2018;Baño-Medina et al, 2019;Pan et al, 2019 , andwind speeds Li, 2019 . Most of these earlier studies focused on obtaining variables at a grid spacing of approximately 10 -50 km by downscaling large-scale variables simulated by general circulation models.…”

Section: Introductionmentioning

confidence: 99%

Bias correction for spatially interpolated daily mean air temperature during winter in eastern Hokkaido using multimodal machine learning

Murakami

Hirota

Shimoda

et al. 2020

J. Agric. Meteorol.

View full text Add to dashboard Cite

Interactions between boundary layer wind and topography form non-uniform air temperature distributions in cold and snow-covered regions. Because of this heterogeneity, spatially interpolated air temperatures sometimes deviate from observed values. To evaluate the quality of spatially interpolated daily mean temperatures T int provided by a 1 km gridded meteorological data service Ohno et al., 2016 , we collected observed temperatures T obs obtained at meteorological observation sites located near farmland in the Tokachi and Okhotsk regions in eastern Hokkaido, Japan in winter October-March and revisited the bias in the interpolated temperatures dT. The root-mean-square error RMSE of T int obtained at 88 sites was 1.16 C, and the absolute median dT values were greater than 1 C at 14 sites. The variance of dT was greater on cold and calm days, suggesting the involvement of radiative cooling and the accumulation of cold air parcels. To correct T int by estimating dT at a given site by considering the formation mechanisms of the temperature distributions, we attempted to develop a multimodal machine learning model that had four predictors: surface and boundary layer meteorological data and topographical and geographical features around each site. To analyze the influence of the spatial extent of the topography and geography around each site, we compared models having these predictors with various sizes of the region of interest ROI. By training the models and applying them to an independent test dataset, it has been shown that bias correction using models with a small topographical ROI 30 30 km or smaller reduced the RMSE. The RMSE of the test dataset decreased by ~0.1 C via the application of a nested model, suggesting the potential usefulness of the presented approach for locally confined meteorological events. However, the biases were increased at several sites by application of the models, thus implying that further improvement is essential for practical use.

show abstract

“…In the first case, in which a large amount of coarse-and fine-grained target data are available, we can predict the fine-grained target data by using a mapping function from coarse-to fine-grained data. The mapping function can be learnt by using various machine learning methods including linear regression models (Hessami et al 2008), neural networks (Cannon 2011;Misra, Sarkar, and Mitra 2017) and support vector machines (Ghosh 2010). Recently, superresolution techniques based on deep neural networks have been applied for refining coarse-grained spatial data (Vandal et al 2017;Vandal et al 2018).…”

Section: Related Workmentioning

confidence: 99%

Refining Coarse-Grained Spatial Data Using Auxiliary Spatial Data Sets with Various Granularities

Tanaka

Iwata

Tanaka

et al. 2019

AAAI

View full text Add to dashboard Cite

We propose a probabilistic model for refining coarse-grained spatial data by utilizing auxiliary spatial data sets. Existing methods require that the spatial granularities of the auxiliary data sets are the same as the desired granularity of target data. The proposed model can effectively make use of auxiliary data sets with various granularities by hierarchically incorporating Gaussian processes. With the proposed model, a distribution for each auxiliary data set on the continuous space is modeled using a Gaussian process, where the representation of uncertainty considers the levels of granularity. The finegrained target data are modeled by another Gaussian process that considers both the spatial correlation and the auxiliary data sets with their uncertainty. We integrate the Gaussian process with a spatial aggregation process that transforms the fine-grained target data into the coarse-grained target data, by which we can infer the fine-grained target Gaussian process from the coarse-grained data. Our model is designed such that the inference of model parameters based on the exact marginal likelihood is possible, in which the variables of finegrained target and auxiliary data are analytically integrated out. Our experiments on real-world spatial data sets demonstrate the effectiveness of the proposed model.

show abstract

Statistical downscaling of precipitation using long short-term memory recurrent neural networks

Cited by 68 publications

References 46 publications

Augmented Convolutional LSTMs for Generation of High-Resolution Climate Change Projections

Augmented Convolutional LSTMs for Generation of High-Resolution Climate Change Projections

Bias correction for spatially interpolated daily mean air temperature during winter in eastern Hokkaido using multimodal machine learning

Refining Coarse-Grained Spatial Data Using Auxiliary Spatial Data Sets with Various Granularities

Contact Info

Product

Resources

About