COLLEGE PARK, Md. – University of Maryland’s Earth System Science Interdisciplinary Center researcher Dr. Xin-Zhong Liang, also a professor in the CMNS-Atmospheric & Oceanic Science Department, released a regional climate model last week called the Climate-Weather Research Forecasting Model (CWRF), a significant extension of the original Weather Research Forecasting Model (WRF).
The model has been ten years in the making by Dr. Liang’s team in collaboration with NOAA and The National Center for Atmospheric Research (NCAR).
Liang began his work updating NCAR’s WRF model a decade ago, but said that he did not think his results were good enough to publish until just this last year.
“There were fundamental issues in the model, so we changed almost every aspect of physics,” he said. He added a climate extension to the WRF model, which initially only included short-term weather forecasts.
The new CWRF model facilitates the use of an optimized physics ensemble approach to improve weather forecast or climate prediction and their impacts on terrestrial hydrology, coastal oceans, crop growth, air quality, water quality and ecosystems.
The model meets users’ needs to simulate climate variability and change in the past and future at regional-local scales and predict consequences on the above mentioned areas.
This means that the model can help give users critical information on the climate for the coming season through the next few years, which can then help determine public policy to the actions of private companies.
Next, the model can predict what the consequences of those actions would really be.
“We can use NOAA data to downscale to regional and local information and then people can use CWRF as a tool to study what happens,” said Liang. “For example, if we change current agriculture into biofuel, what are the consequences in hydrology or air quality?”
The model is able to predict these consequences because it incorporates surface boundary conditions, something that many people in the modeling community pay insufficient attention to, said Liang.
“The data for this is very rare, so we worked very hard to get this part, including things like where a river bed is, where the stream flows,” Liang said. “So we had to use information from satellites and samples and combine all of the information.”
This data will give you the ‘What will happen?’ answer, he said. For example, if a policy maker decided to regulate emissions in the U.S. in order to improve the air quality, the air above the specified area may actually be coming from Canada or Mexico, making his legislation worthless.
“Before you take off with expensive engineering, you look at the model,” said Liang. “This can give you guidance before action taking. If you know the consequences, you may ask how you are going to control and mitigate the issues. You can use the model and ask if the strategy you plan to do is effective or not.”
While confident today that the model is a success, Liang spent the majority of the last decade making sure that everything would run smoothly.
“We wanted the model to be able to capture extreme events and anomalies and repeat them,” Liang said. “We spent a lot of time checking that they would predict these big events.”
Liang and his team spent time on retrospective testing, as well as testing in real time. They would take data from NOAA’s real time predictions, pass it through CWRF, and see if their results were an improvement over NOAA’s predictions. They found that it added substantial values, especially for precipitation, which is difficult to deal with in world climate predictions, said Liang.
The model is in high demand in the climate prediction community, and was already been heavily cited since its development and before public release.
Since the model’s release to the public on Friday afternoon, over 150 people have registered for its use. Still, the team faces some technical issues, like having too much data on their server or installation issues. They also hope to add a forum for users to discuss applications for the model with each other.
The model also has some slight prediction issues, like producing too much rain in Canada and in the Gulf of Mexico, said Liang, but the team knows why this is happening and how to fix it.
They’re continuing to make the model better suited to user needs as well. With huge amounts of data sets and physics schemes available, the next step is to develop an optimized modeling system, said Liang.
One way of doing this is having users select their use for the model when they register, choosing from a list of options including research, education, or private sector options.
If someone registers for any of the research options, which include model development and climate prediction, then Liang and his team will be able to contact them to collaborate on developments.
Dr. Liang and his team will publish a paper on CWRF in the Bulletin of the American Meteorological Society in the next month, where it is expected to generate large responses within the scientific and climate forecasting community, said Liang.
“The release will be tremendously useful for the community in climate study.”
CWRF Model Home: http://cwrf.umd.edu
Climate Information: Responding to User Needs: http://www.climateneeds.umd.edu