WRF Administration

Staff Directory

Accomplishments

DTC Events, Announcements
& Past Events

WRF Code in the DTC

WRF website - 
   ARW User's Page
   NMM User's Page
   MET User's Page

Data Assimilation Testbed Center

Announcement of Opportunity

Announcement of Opportunity

with the Weather Research and Forecasting Model (WRF)

Developmental Testbed Center

The WRF Developmental Testbed Center (DTC) Visitor Program is pleased to announce that approximately seven new visitor appointments will be available for the year beginning June 15, 2007. These appointments will enable visitors to work with the DTC in testing new techniques, models, and model components for numerical weather prediction. The goal is to provide the operational weather prediction centers, NCEP, AFWA and FNMOC, with options for near-term advances in operational weather forecasting and to provide researchers with numerical weather prediction (NWP) codes that represent the latest advances in the technology.

Successful applicants will be offered up to one month of salary compensation, travel and per diem. The visitors are expected to spend one month with the DTC in Boulder, Colorado, in Monterey, California, or at one of the operational centers. This one-month visit can be distributed over several weeks during a one-year period. Access to DTC computational resources will enable significant portions of the visitor's project to be conducted from their home institution.

1.0 The WRF Developmental Testbed Center (DTC)

The WRF program includes plans for the rapid and direct transfer of new research results into the NWP process of the National Weather Service (NWS) and other operational NWP centers. It also includes plans for providing the research community with the latest NWP technology and access to the operational models for various research projects. The WRF effort embodies the concept of the operational and research communities working jointly toward development of next generation NWP capabilities that will allow, as new techniques are developed in the research community, the most promising results to be rapidly and efficiently transferred to operations.

The DTC is a distributed facility in which the operational and research communities will work closely together in developing and testing the next generation numerical forecast systems. In the development process, researchers will be invited to work within the DTC with Center personnel and with members of the operational community and for members of the operational community to work with researchers to demonstrate the promise of new techniques in NWP. In addition, members of both the operational and research community will be able to evaluate the current operational models through retrospective analysis and diagnosis of their strengths and weaknesses. A key objective of the DTC is to offer to the research community an environment that is functionally similar to that used in operations to test and evaluate new NWP methods, without interfering with actual day-to-day operations of the operational centers.

The DTC includes components in Boulder , Colorado and at the Naval Research Laboratory in Monterey , California ; the component in Boulder is referred to as the Boulder DTC and the component in Monterey , California is referred to as the NRL DTC. In addition, the Boulder DTC is composed of components at NOAA's Global Systems Division (GSD DTC) and NCAR (NCAR DTC). The visitor could be associated with any of the three components; those visiting Boulder would be primarily located at NCAR but would interact with DTC personnel at both NCAR and GSD. In addition, since a goal of the DTC is to transition research into operations, a visitor to the DTC could also be associated with any of the operational centers such as EMC at NCEP.

2.0 The WRF Code system

The WRF code systems will consist of three formal levels: Contributed Code, Reference Code, and Operational Code. Management of the Contributed and Reference Codes, as well as how researchers can contribute to these codes, is currently being discussed; results of these discussions will be made public when available. Below we describe in general terms each of the three code categories.

2.1 Contributed Code

Contributed Code will be the most informal class of code in the WRF system. The main requirements for inclusion in this category are that t he code is compliant with the WRF coding infrastructure and that the technique addresses a potential operational weather forecast need or represents an advancement in NWP technology. Generally the codes' authors will maintain these codes, but it is expected that they will be made available to the community through a code repository maintained by NCAR/MMM and the DTC. Visitors to the DTC could conduct tests on components of the Contributed Codes.

2.2 Reference Code

This set of code will be the heart of the WRF system. Reference Code will consist of carefully selected tested codes and will have available results of tests conducted by the DTC, including a limited set of verification statistics. It will also contain the configuration of the latest WRF codes being run operationally (updated on the order of twice a year) . This code will be fully maintained by the DTC and will be made available to the community. Users of the DTC will work primarily with the Reference Code.

The initial Reference Code contains multiple WRF cores, physics options, initialization systems, post-processing systems and a verification system. This will be expanded in the future to contain the components to replicate the hurricane WRF system, data assimilation systems, an advanced verification system. Eventually it may contain the operational global forecast models. Currently the WRF Reference Code consists of two dynamical cores; the Non- Hydrostatic Mesoscale Model (NMM) and the Advanced Research WRF (ARW). There are two complete physics packages, one from NCEP that was designed to work with the NMM core and one from NCAR that is part of the ARW core and the NCEP verification system. Each of these packages can be interchanged between the cores as can components of the physics. Where applicable, visitors are encouraged to work with both WRF cores.

2.3 Operational Codes

The Operational Codes are the fully-hardened, fully-tested codes that are being run operationally at the various operational centers. This level of code will be maintained by the Operational Testbed Centers (OTC) at the operational centers but made available to the community through the DTC. These Operational Codes are a subset of the Reference Code and, as such, are also candidates for researchers' attention at the DTC.  A goal of producing incremental upgrades to address weaknesses in existing Operational components will be viewed as a valid topic for DTC applicants.

3.0 How to Respond to this Announcement

In section 4.0, possible projects that are of interest to the DTC are outlined. These are general, and proposals for participation in the visitor program should provide details on the specific work that the visitor would conduct with the DTC. This proposal should be described in a document no longer than 5 pages. The submitted material should include a brief one-page summary of the project, a C.V. of no more than 2 pages, and a budget for 1-2 months of salary for the PI and travel costs. As noted above, it is expected that the visitor will spend one month in residence at one of the distributed DTC sites and that the total duration of the project can continue for one year. It is expected that the visitor will be able to continue the work from his or her own institution using DTC computational resources.

Proposals in response to this announcement should be sent by April 30, 2007 to:

Pam Johnson
NCAR/DTC
P.O. Box 3000
Boulder , Colorado 80307

Express mail address: 3450 Mitchell Lane , Boulder , CO , 80301

Or send electronically to: johnsonp@ucar.edu

4.0 Possible Visitor Projects with the WRF DTC

This is a general announcement of an opportunity to work with the WRF DTC to test existing WRF-based NWP systems in order to assess where they are deficient, and to assess new NWP technology that shows promise of improving numerical weather prediction within the next five years. This could include testing new physics parameterization components, optimizing physics packages, comparing dynamic cores, alternative verification approaches, data assimilation systems, as well as investigations of the impacts of resolution and the tradeoffs between an ensemble versus a deterministic approach.

Some more specific suggested topics that would receive special consideration include:

1. Add the ability for the NMM core to test ideal cases such as what conditions in the model dictate flow splitting and going around a barrier as opposed to rising up and flowing over it.  NCEP has noted some interesting (and different) behavior between the cores for the Black Hills of South Dakota as an example. This task may address the more general topic of behavior of mountain induced flows in the WRF models.
   

2. There are a number of new PBL schemes that are being tested in the community that might be a welcome addition to the WRF system. A possible project therefore is to add a new boundary layer parameterization scheme to the WRF physics options, and test and compare this scheme with existing schemes both in idealized one-dimensional cases, and in full NWP-type simulations. This testing could be extended to seasonal verifications for statistical evaluation, using the DTC datasets, and collaborating with DTC staff.
   

3. In the next couple of years the DTC will be enabling an ensemble capability to provide to the user community. The DTC would welcome visitors who want to assist us in assembling an end-to-end ensemble modeling system including components to generate an ensemble and post processing components specific to ensemble systems.
   

4. How much smoothing is "necessary" for the underlying terrain in the WRF cores?  This depends on the inherent model smoothing and so may be core- dependent. The core comparisons conducted by the DTC made some attempt to make the underlying terrain as identical as possible in both cores but this is difficult because of the use of different grids. A possible project therefore is to make the terrain as identical as possible and conduct some retests to determine if the core differences get smaller or remain the same. Experiments can also be proposed for other tests of the WRF models to their sensitivity to terrain treatments such as form drag and mountain blocking. The core tests are being extended this summer to longer time intervals and higher resolutions. A visitor interested in this project can work with the DTC in evaluating or perhaps further extending these tests.
   
   A related topic would be for a visitor to work closely with DTC staff currently investigating the ability of various WRF model configurations at high resolution to properly simulate observed mountain wave structure and dynamics. Particular interest is focused on use of special T-REX datasets to verify model forecasts, and also the development and testing of different numerical approaches for treating the vertical propagation of wave energy and damping of reflective gravity waves from the model's top boundary.
   

5. What is the value-added of including two-way interactive nesting verses one-way nesting?  These tests should include both WRF cores using standard verification scores for all nests especially the coarsest parental nest to see if there is positive feedback to the larger scales.  Answers to this question are of particular interest to the operational centers since in the future operational global and mesoscale models will be run concurrently and there is the choice of whether to build two-way nesting (so the mesoscale model can feedback into global and therefore the two models must be run together) or just a simple one-way coupler (which may allow more flexible scheduling for the operational model computations).
   

6. A variety of projects focusing on advanced verification methods would be of interest. Some examples include the following:
   
   a. Investigate verification approaches that are more appropriate for providing model diagnostics than many traditional approaches – that is, approaches that provide information about particular attributes of model error that can lead to a diagnosis of needed improvements in the model.
   
   b. Apply verification approaches that allow incorporation of observational uncertainty in model evaluations, and are able to express this uncertainty in the resulting verification measures. Demonstrate how this information can be separated from other sources of uncertainty (e.g., sampling variations) associated with estimates of verification measures. For example, it would be of interest to investigate the impacts of variations among different types of analyses (e.g., for precipitation) used as verification observations on variations in traditional and non-traditional verification measures.
   
   c. Apply new spatial verification methods for evaluation of ensemble forecasts.
   
   d. Incorporate the time dimension in a spatial verification approach so that timing errors can be diagnosed, and application of this approach to a variety of WRF cases.
   

5.0 Proposal Evaluation Process

The proposals submitted in response to this announcement are subject to both external and internal review. The external review will be conducted by an advisory board appointed by the DTC, which will consist of atmospheric scientists from government labs, operational centers, and academic institutions. The DTC Director, in consultation with the DTC Executive Committee, will make the final selections based on the review by the Advisory Board.