Improving satellite measurements of clouds and precipitation using machine learning
Doktorsavhandling, 2022

Observing and measuring clouds and precipitation is essential for climate science, meteorology, and an increasing range of societal and economic activities. This importance is due to the role of clouds and precipitation in the hydrological cycle and the weather and climate of the Earth. Furthermore, patterns of cloudiness and precipitation interact across continental scales and are highly variable in both space and time. Therefore their study and monitoring require observations with global coverage and high temporal resolution, which currently can only be provided by satellite observations.

Inferring properties of clouds or precipitation from satellite observations is a non-trivial task. Due to the limited information content of the observations and the complex physics of the atmosphere, such retrievals are endowed with significant uncertainties. Traditional methods to perform these retrievals trade-off processing speed against accuracy and the ability to characterize the uncertainties in their predictions.

This thesis develops and evaluates two neural-network-based methods for performing retrievals of hydrometeors, i.e., clouds and precipitation, that are capable of providing accurate predictions of the retrieval uncertainty. The practicality and benefits of the proposed methods are demonstrated using three real-world retrieval applications of cloud properties and precipitation. The demonstrated benefits of these methods over traditional retrieval methods led to the adoption of one of the algorithms for operational use at the European Organisation for the Exploitation of Meteorological Satellites. The two other algorithms are planned to be integrated into the operational processing at the Brazilian National Institute for Space Research, as well as the processing of observations from the Global Precipitation Measurement, a joint satellite mission by NASA and the Japanese Aerospace Exploration Agency.

The principal advantage of the proposed methods is their simplicity and computational efficiency. A minor modification of the architecture and training of conventional neural networks is sufficient to capture the dominant source of uncertainty for remote sensing retrievals. As shown in this thesis, deep neural networks can significantly improve the accuracy of satellite retrievals of hydrometeors. With the proposed methods, the benefits of modern neural network architectures can be combined with reliable uncertainty estimates, which are required to improve the characterization of the observed hydrometeors.

clouds

remote sensing

machine learning

precipitation

EB, Edit
Opponent: Dr. Peter Dueben, Coordinator of Machine Learning and AI activities at the European Centre for Medium Range Weather Forecasts

Författare

Simon Pfreundschuh

Chalmers, Rymd-, geo- och miljövetenskap, Geovetenskap och fjärranalys

Can machine learning correct microwave humidity radiances for the influence of clouds?

Atmospheric Measurement Techniques,; Vol. 14(2021)p. 2957-2979

Artikel i vetenskaplig tidskrift

A neural network approach to estimating a posteriori distributions of Bayesian retrieval problems

Atmospheric Measurement Techniques,; Vol. 11(2018)p. 4627-4643

Artikel i vetenskaplig tidskrift

S. Pfreundschuh, P. J. Brown, C. D. Kummerow, P. Eriksson, and T. Norrestad. GPROF-NN: A neural network based implementation of the God- dard Profiling Algorithm. Atmospheric Measurement Techniques Discussions, 2022:1–37, 2022a. doi: 10.5194/amt-2022-15

S. Pfreundschuh, C. Guilloteau, P. J. Brown, C. D. Kummerow, and P. Eriksson. Evolution of the GPROF passive microwave precipitation retrievals evaluated against ground radar measurements over the continental US and the Pacific Ocean. In preparation, 2022c

S. Pfreundschuh, I. Ingemarsson, P. Eriksson, D. A. Vila, and A. J. P. Calheiros. An improved near real-time precipitation retrieval for Brazil. EGU- sphere, 2022:1–30, 2022d. doi: 10.5194/egusphere-2022-78

The Earth's climate is changing and with it the distribution and frequency of clouds and rain. As a result, heavy rain and droughts are expected to increase. Humans need to detect and monitor these changes in order to be able to adapt to them. Satellite observations are essential for this since they are the only way such measurements can be obtained on a global scale.

This thesis employs methods from artificial intelligence to improve satellite measurements of clouds and rain. These measurements are difficult because the atmosphere is highly complex and the information provided by the satellite is limited. Providing reliable measurements thus requires quantifying the associated uncertainties. This thesis proposes a way to employ novel machine learning methods in a way that allows reliable quantification of the uncertainties.

Using three different satellite applications, the potential of the proposed methods is demonstrated. They yield significant improvements compared to current methods and improve the handling of uncertainties in the measurements. Because of these advantages, the methods are already used by the European Organisation for the Exploitation of Meteorological Satellites and planned to be used for the processing of observations from the Global Precipitation Measurement, a joint satellite mission by NASA and the Japanese Aerospace Exploration Agency.

Robust Estimation of Global precipitation using Neural networks (REGN)

Rymdstyrelsen (154/19), 2020-01-01 -- 2023-12-31.

Ämneskategorier

Matematik

Data- och informationsvetenskap

Geovetenskap och miljövetenskap

ISBN

978-91-7905-657-5

Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5123

Utgivare

Chalmers

EB, Edit

Online

Opponent: Dr. Peter Dueben, Coordinator of Machine Learning and AI activities at the European Centre for Medium Range Weather Forecasts

Mer information

Senast uppdaterat

2022-05-06