Browse Prior Art Database

Analytics Platform for Numerical Weather Prediction Disclosure Number: IPCOM000236327D
Publication Date: 2014-Apr-21
Document File: 4 page(s) / 90K

Publishing Venue

The Prior Art Database


Disclosed is a networked architecture of data-centric middleware components in a High Performance Computing (HPC) environment to provide a scalable and robust data analytics platform suitable for the development of Numerical Weather Prediction (NWP) models.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 4

Analytics Platform for Numerical Weather Prediction

I. Introduction

Modern-day forecasting relies on Numerical Weather Prediction (NWP) models running in High Performance Computing (HPC) servers. By using NWP models it is possible to accurately predict severe weather events by using realistic computer models of clouds, precipitation, and convective processes. NWP models generate data that could be useful for disaster preparation and decision making. As more computing power becomes available to run NWP models, scientists are now observing that performing subsequent data processing and analysis are increasingly becoming the major bottlenecks to effectively utilizing the HPC systems. It is now more important to consider data management issues inherent in large-scale, data-intensive computing. Decision support systems that facilitate data exploration and analysis can significantly reduce the time and effort to perform weather forecasts using NWP models.

Disclosed is a software infrastructure that uses a unique combination of middleware components to address two fundamental challenges in modern-day NWP systems:
(1) to manage and process large amounts of data arriving in time-sensitive streams from remote sensors and weather models; and (2) to significantly reduce data analysis cycles so that scientists can make timely decisions and improvements to numerical weather prediction models.

II. NWP Model Development Workflow

Figure 1 is a data flow diagram showing the typical processes involved in running and analyzing NWP models.

The basic idea of NWP is to sample the state of the atmosphere at a given time and use the equations of fluid dynamics and thermodynamics to estimate the state of the atmosphere at some future time. NWP models are initialized using real-time environmental data coming from remote sensors via a process called data assimilation. An objectives analysis is applied to the raw sensor data to perform quality control and obtain values as input to the model. The processed data are used by the model as the starting point for a forecast.

Forecast verification is used to assess the accuracy of the model output. This step is important especially in identifying areas of improvement in the model. In weather forecasting, skill scores are used to gauge the accuracy of forecasts generated by NWP models. Skill scores are computed based on how far the weather forecast is from the actual observations. By analyzing the performance of the model, scientists can gain insights on how to further improve the model.

Figure 1. Data Flow Diagram of NWP Model Development.


Page 02 of 4

III. Design

Figure 2 shows a schematic diagram of the architecture labeled with the different mid...