Award Abstract # 1331915
CyberSEES:Type 2: Precipitation Estimation from Multi-Source Information using Advanced Machine Learning

NSF Org: CCF
Division of Computing and Communication Foundations
Recipient: UNIVERSITY OF CALIFORNIA IRVINE
Initial Amendment Date: September 10, 2013
Latest Amendment Date: September 10, 2013
Award Number: 1331915
Award Instrument: Standard Grant
Program Manager: Eva Zanzerkia
CCF
 Division of Computing and Communication Foundations
CSE
 Directorate for Computer and Information Science and Engineering
Start Date: September 15, 2013
End Date: August 31, 2018 (Estimated)
Total Intended Award Amount: $1,060,000.00
Total Awarded Amount to Date: $1,060,000.00
Funds Obligated to Date: FY 2013 = $1,060,000.00
History of Investigator:
  • Soroosh Sorooshian (Principal Investigator)
    soroosh@uci.edu
  • Xiaogang Gao (Co-Principal Investigator)
  • Kuolin Hsu (Co-Principal Investigator)
  • Alexander Ihler (Co-Principal Investigator)
Recipient Sponsored Research Office: University of California-Irvine
160 ALDRICH HALL
IRVINE
CA  US  92697-0001
(949)824-7295
Sponsor Congressional District: 47
Primary Place of Performance: University of California-Irvine
CA  US  92697-2700
Primary Place of Performance
Congressional District:
47
Unique Entity Identifier (UEI): MJC5FCYQTPE6
Parent UEI: MJC5FCYQTPE6
NSF Program(s): CyberSEES
Primary Program Source: 01001314DB NSF RESEARCH & RELATED ACTIVIT
Program Reference Code(s):
Program Element Code(s): 821100
Award Agency Code: 4900
Fund Agency Code: 4900
Assistance Listing Number(s): 47.070

ABSTRACT

This project will develop a cyber-enabled, data-driven modeling system that will use the vast amount of earth and environmental observational data to estimate precipitation, with the goal to further freshwater resource sustainability. One of the major objectives of this research is to innovatively apply advanced machine learning techniques to predict complex natural phenomena. Identification of current machine learning algorithms' representational and computational limitations when applied to earth and environmental data that are required for accurately estimating precipitation will also be explored. Specifically, this project will explore image-based feature extraction techniques and so-called "deep belief" models for interpreting important features within weather and climate data to estimate and predict precipitation. Features to be investigated include appearance, texture, shape, dynamics, and regional weather and climate signatures. The project will use data from ground-based radar, satellite-based sensors, and Numerical Weather Prediction (NWP) models, as well as physical characteristics from land surface datasets. Image-based extraction techniques will produce feature maps, which in turn will be used as input into a Deep Boltzmann Machine (DBM) model, a modern version of neural networks. A DBM represents a probability model over a collection of visible units and hidden units and produces as output a target variable (in this case, precipitation). This model is particularly suitable for the proposed modeling approach and for large-scale data. Once this system is developed, we will utilize a wide variety of earth science data to provide accurate, high-resolution global precipitation estimates. Verification methods to be used for evaluating precipitation estimates will include general statistics, precipitation intensity distribution, and regional analysis methods used by the atmospheric and hydrological sciences. The project will focus on verification over the United States, where high-resolution ground-based radar data are available.

In recent decades, extreme flooding and droughts have become more frequent and severe. Changing patterns in precipitation, which are attributed to climate variability, are responsible for these hydrologic extremes and contribute to uncertainties in freshwater resource management and planning. In the face of the planet's growing population and stresses on water resources, it is important to minimize these uncertainties and the social impacts of these natural hazards. These goals can only be accomplished through accurate precipitation measurements and forecasts. Satellite platforms and advanced numerical models produce massive amounts of global high temporal and spatial resolution data that can be used for this purpose, but analyzing these data remains a challenge. Recent innovations in computational sciences and machine learning have extended our capability to harvest from remotely-sensed data critical information that is essential to understanding cloud-precipitation systems. This project will adapt and improve satellite-based precipitation estimation algorithms by using computational science, statistical modeling techniques (machine learning), remote-sensing observations, and numerical models to develop cyber-enabled modeling systems that can effectively analyze the massive amount of observational data to improve the global estimation of precipitation at high spatial and temporal resolutions.

PUBLICATIONS PRODUCED AS A RESULT OF THIS RESEARCH

Note:  When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

Akbari Asanjan, A.,T. Yang, K. Hsu, S. Sorooshian, J. Lin, and Q. Peng "Short-term Precipitation Forecast Based on the PERSIANN System and the Long Short-Term Memory (LSTM) Deep Learning Algorithm" Journal of Geophysical Research ? Atmospheres , 2018 10.1029/2018JD028375
Nguyen, P., A. Thorstensen, S. Sorooshian, K. Hsu, A. AghaKouchak, H. Ashouri, H. Tran, and D. Braithwaite "Global Precipitation Trend across Spatial Scales Using Satellite Observations" Bulletin of American Meteorological Society , 2018 10.1175/BAMS-D-17-0065.1
Nguyen, P., S. Sorooshian, A. Thorstensen, H. Tran, P. Huynh, T. Pham, H. Ashouri, K. Hsu, A. AghaKouchak, and D. Braithwaite "Exploring Trends Through ?RainSphere? Research Data Transformed into Public Knowledge" Bulletin of American Meteorological Society , 2018 10.1175/BAMS-D-16-0036.1
Ping, W. and Ihler, A. "Belief Propagation in Conditional RBMs for Structured Prediction" Proceedings of the 20th International Conference on Artificial Intelligence and Statistics , v.54 , 2017 , p.1141
Ping, W. and Ihler, A. "Learning Infinite RBMs with Frank-Wolfe" Advances in Neural Information Processing Systems 29 , 2016 , p.3063
Tao, Y., Gao, X., Hsu, K., Sorooshian, S. and A. Ihler "A Deep Neural Network Modeling Framework to Reduce Bias in Satellite Precipitation Products" Journal of Hydrometeorology , v.17 , 2016 http://dx.doi.org/10.1175/JHM-D-15-0075.1
Tao, Y., k. Hsu, A. Ihler, X. Gao, and S. Sorooshian "A Two-Stage Deep Neural Network Framework for Precipitation Estimation from Bi-spectral Satellite Information" Journal of Hydrometeorology , v.19 , 2018 10.1175/JHM-D-17-0077.1
Tao, Y., X. Gao, A. Ihler, S. Sorooshian, and K. Hsu "Precipitation Identification with Bispectral Satellite Information Using Deep Learning Approaches" Journal of Hydrometeorology , v.18 , 2017 , p.1271 10.1175/JHM-D-16-0176.1
Tao, Y., X. Gao, K. Hsu, S. Sorooshian, and A. Ihler "A Deep Neural Network Modeling Framework to Reduce Bias in Satellite Precipitation Product" Journal of Hydrometeorology , v.17 , 2016 , p.931 10.1175/JHM-D-15-0075.1

PROJECT OUTCOMES REPORT

Disclaimer

This Project Outcomes Report for the General Public is displayed verbatim as submitted by the Principal Investigator (PI) for this award. Any opinions, findings, and conclusions or recommendations expressed in this Report are those of the PI and do not necessarily reflect the views of the National Science Foundation; NSF has not approved or endorsed its content.

This project integrates state-of-the-art computational scientific methods to improve the performance and accuracy of global precipitation estimation. Meteorological data from satellite sensors and Deep Neural Networks (DNNs) were used for precipitation estimation in two steps.  DNN was trained to detect rain/no-rain (R/NR) binary classification, and followed to estimate rainfall rates. This study also extended from rainfall estimation to the short-term rainfall forecasting up to 6 hours using a Long Short-Term Memory (LSTM) Recurrent Neural Network. The LSTM shows effective for rainfall forecasts while compared with other conventional methods, such as persistency method, optical flow extension, and the numerical model.   

This project developed cyber-enabled data driven modeling systems that use the vast amount of earth and environmental observational data to estimate precipitation. Our study explored deep learning neural networks for interpretation of spatial and temporal weather related information for the precipitation estimation and forecast. The development including using a large collection of multi-spectral images from multi-satellites at high spatial and temporal resolutions. High-resolution of ground based radar data were used in the calibration and validation. The developed system has potential to be used to a wide variety of earth science data to provide accurate, high-resolution resolution precipitation estimates from regional to global scales. This proposal reflects the synergistic collaboration between hydrometeorology and machine learning experts that together will provide a unique opportunity to develop innovative methods and new knowledge within both disciplines. 

Extreme precipitation events causing flooding (due to excessive precipitation) and droughts (due to long-time scarcity of precipitation) have become more frequent and severe. Climate variability and change are major sources of uncertainty in freshwater resource management and planning, and society needs reliable water resources. Accurate measurements and forecasts of precipitation are critical for the preparation and planning by society to minimize the impact of these natural hazards. Satellite platforms produce massive amounts of global high temporal and spatial resolution data. Recent innovation in computational and Machine Learning sciences have extended our capability to harvest critical knowledge and information essential to cloud-precipitation systems from the vast amounts of remotely-sensed data available. This project adapts and improves satellite-based precipitation estimation algorithms by using computational science, statistical modeling techniques, remote-sensing observations and numerical models to develop cyber-enabled modeling systems to harness the massive amount of observational data to improve the estimation of precipitation at high spatial and temporal resolution.

 


Last Modified: 12/04/2018
Modified by: Kuolin Hsu

Please report errors in award information by writing to: awardsearch@nsf.gov.

Print this page

Back to Top of page