SR/SSD 98-35


Technical Attachment

Forecaster Interactions with Global Hydrology and Climate Center

Tom Bradshaw

NWSFO Birmingham

Forecasters at NWSFO Birmingham and NWSO Nashville are collaborating with researchers at the Global Hydrology and Climate Center (GHCC) in Huntsville, Alabama, to provide real-time evaluations of temperature and precipitation forecasts from an experimental version of the MM5 mesoscale model. The GHCC comprises atmospheric scientists from NASA's Marshall Space Flight Center and the Department of Atmospheric Sciences at the University of Alabama in Huntsville. The following is a summary provided by Tom Bradshaw at NWSFO Birmingham, describing in more detail the nature of this interaction. Ed.

With support provided by the U.S. Weather Research Program, Dr. Bill Lapenta of NASA/MSFC and Dr. Richard McNider of UAH are conducting a research project involving the assimilation of GOES skin temperature and albedo products into the MM5 mesoscale model. The goal of this project is to improve the parameterization of the surface energy budget within the model, which in turn should lead to improved QPF and temperature forecasts during the warm season. Last spring Dr. Lapenta traveled to NWSFO Birmingham to seek our help in evaluating the MM5 products on a real-time basis.

Since early July the GHCC has been providing a variety of real-time MM5

forecast fields on their Web page. The site is:

GHCC runs their model twice each day on two domains, a Southeast domain and a CONUS domain, posting their 48-hr output to their Web site before noon and midnight. Their site is fast, extremely reliable, and easy to navigate. It is one of the most outstanding ones we've seen, in fact, for modeling activities - whether mesoscale or otherwise.

NWSFO Birmingham has been conducting a daily evaluation of the MM5 QPF vs. the Eta and NGM products, and we post results on a Web site on our office Intranet for access by forecasters. The MM5 QPF output has some quirks, especially during synoptically disturbed episodes, but during quiescent conditions it is more than a match for the Eta and NGM. More importantly, during the past few weeks, the Birmingham and Nashville WFOs have been forwarding comments on the accuracy of these products directly to Dr. Lapenta.

Summary of the July 31, 1998 Meeting at GHCC.

At the end of July, Ron Murphy and the author from NWSFO Birmingham visited GHCC and met with participants from the GHCC MM5/GOES Assimilation Project, including Dr. Bill Lapenta (NASA), Dr. Richard McNider (UAH), Dr. Gary Jedlovec (NASA), Dr. Qi Mao (UAH), and Mr. Ron Suggs (NASA). The MM5 researchers expressed appreciation for the input we've provided them, and in return we let them know how much we appreciate the opportunity to interact with them and access the data. Both sides expressed an interest in closer cooperation. In September the GOES assimilation process will begin, so this will give us a chance to see how much improvement this technique will have on the model. I anticipate that we will be providing quite a bit of additional input on the quality of the data through the coming fall months.

At the meeting Dr. Lapenta provided a brief overview of the project's objectives and methodologies. He reiterated how useful our comments and those of Henry Steigerwaldt (SOO, NWSO Nashville) were in determining the initial reliability of the MM5 products. He recapped the project proposal he had already provided our office, and in particular discussed the motivation behind assimilating GOES skin temperature, surface albedo, and insolation time tendency information into a mesoscale model to provide a more realistic representation of the model's surface energy budget. The crux of the project centers on the fact that surface flux characteristics are extremely difficult to model adequately. Dr. McNider has developed a scheme whereby these surface processes can be parameterized by dynamically assimilating GOES data into the early forecast hours of the model cycle.

Basically, the model is started and allowed to run for a couple of hours. At this point, GOES-derived skin temperature, albedo and insolation tendency fields are collected and compared with those forecast by the MM5. The GOES fields are used to "nudge" the MM5 forecast temperature and moisture fields closer to the actual state of the surface energy budget as it evolves through the morning hours. This will hopefully lead to improved forecasts of temperature and precipitation in the later stages of the model run.

Dr. Lapenta also gave an update on the status of the GHCC MM5 real-time product cycle, and presented a project timeline through the year 2000. Currently, the model is running twice a day, without the GOES assimilation. He hopes to begin the assimilation process sometime in September. Additionally, he hopes to run the model operationally through the winter months, pending funding.

Mr. Suggs next provided some more detailed background on the GOES-derived products and how they can be assimilated into a mesoscale model. In particular, he presented some results from the Atmospheric Radiation Program (ARM) in Oklahoma, which demonstrated the improvement in model predicted temperature and dewpoint which could be realized through the dynamic assimilation of GOES insolation data.

Dr Mao presented some findings from a study which is being conducted jointly between UAH and the Tennessee Valley Authority. This study involves the use of the Regional Spectral Model (RSM) centered over the TVA service area, focusing on the ability of this model to provide improved temperature and QPF forecasts which can enhance the ability of TVA to anticipate power load and reservoir control requirements in the 0- to 72-hr time frame. Dr. Mao has developed an interesting procedure for verifying temperature and QPF forecasts from the RSM which hopefully can be utilized for the current MM5 project. He has also developed a new technique for generating MOS-like temperature products from model data. The advantage of this technique is that it doesn't require the model to be "frozen", as is required with the current operational MOS technique employed by TDL.

After these presentations were completed, I discussed some of our activities at NWSFO Birmingham, including our ongoing efforts to compare daily Eta, NGM and MM5 QPF products and observed precipitation products. I displayed some examples of our Intranet web page that displays the data for the benefit of our forecast staff. I also discussed some of my efforts to compare the Eta and MM5 QPF products over the Black Warrior River Basin.

At this point, there was an extended general discussion regarding the overall project, and the quality of the MM5 QPF and temperature data to this point. I raised some of the same issues that had been mentioned via e-mail earlier in the month, such as the problem of poor precipitation spin-up due to the cold start of the model. In reply, Dr. Lapenta described some ongoing efforts to mitigate this problem. I also raised a number of questions regarding the exact process by which the GOES data was to be incorporated in the model. For instance, what impact will cloud-induced spatial and temporal irregularities in the GOES data have on the assimilation scheme? At what temporal resolution will the GOES products be collected and injected into the model? How soon will the model output be posted to the web site? In addition, we discussed some of the positives and negatives of the GHCC web page, as well as the positive attributes of other mesoscale model sites, including the CAPS, NSSL, and FSU web sites.

The members of the project also had a number of questions for Ron and I. For example, Dr. McNider and Dr. Jedlovec asked us how our office utilizes mesoscale model data in our forecast process. Specifically, they wanted to know what types of forecasts we issue, and the means by which we disseminate these products to our users. It was interesting to discover how much (or little) knowledge of the NWS some of these research scientists had. I believe they left the meeting with a better understanding of how we operate, and how we might be able to help them in the future. In particular, we found during the meeting that Dr. Mao had had a difficult time tracking down TDL as the source for the development of the NGM MOS guidance. Clearly, in the future we can help them locate sources of technical assistance from within the NWS or NOAA.

Following the meeting, Dr. Lapenta gave us a quick tour of the GHCC facilities, and he demonstrated show of some of the computer hardware/software he uses to post-process the model output.

The meeting went extremely well from our perspective, and I believe we impressed the GHCC researchers by the degree to which we've utilized their data to this point, and the amount of input we've already provided to this project. We continue to lead all NOAA sites in the number of hits on the GHCC web page. I expressed our willingness to provide as much additional assistance as possible during the next two years. In return, I believe the researchers would like to increase the bond which has developed between our respective organizations. This might be accomplished through future joint projects, perhaps involving COMET partners funding. We look forward to future collaboration.