Western U.S. Testbed Year 1 Report

Western U.S. Testbed Year 1 Report

July, 2006

Authors: A. Wood, T. Pagano, K. Werner, F. Weber

1. Project Purpose

This test-bed is intended to facilitate the comparison and evaluation of practical ensemble forecast related methods that are viable in an operational setting, and address current operational forecast difficulties that are particular to the western U.S. and British Columbia. Priorities include:

• Development and testing of strategies for snow assimilation, and work yielding insight into the effects of snow assimilationon forecast uncertainty and bias.
• Collaborative research focusing on hydrologic model error reduction and estimation via calibration approaches.
• Climate forecast related research, particularly where snow is less important than rain.

The test-bed leaders offered to make available their datasets and models related to those methods for use in comparative evaluations (retrospective, and if possible, in real-time), for the following basins: Mica Dam (BC), Feather R. (CA), Yakima R. (WA), Salmon R. (ID), Gunnison R. (CO), Upper Klamath R. (OR)

2. Accomplishments during the past year

Frank Weber produced a paper describing a framework for hydrologic forecast verification, and used it to assess BC Hydro forecasts for two basins (Stave, and Columbia R. upstream of Mica Dam, BC). In support of test-bed related research, the following data were compiled for the Columbia River upstream of Mica Dam:

• an archive of probabilistic water supply forecasts (1980-2005)
• natural flow data(1980-2005)

Kevin Werner has initiated an effort (involving Andy Wood and Tom Pagano) to compare retrospectiveand “real time” seasonal flow forecasts for 9 basins in the Western US, with forecasts being drawn from NWSRFS, VIC, statistical and other methods. It is expected that this initial set of methods and basins will later be expanded. All results and data are being made available via the web and linked to the test-bed data website (given below).

Tom Pagano led several efforts to improve the availability of operational data from the National Water and Climate Center that will be useful for test-bed related research efforts – in particular for establishing baselines of forecast performance and for forecast verification.

• an archive of probabilistic water supply forecasts (1993-current): ftp://ftp.wcc.nrcs.usda.gov/data/water/forecast/forecast_bounds_byyear/
• deterministic outlooks (1922-2003): ftp://ftp.wcc.nrcs.usda.gov/data/water/forecast/forecast_archive/
• natural flow data for all forecast locations: ftp://ftp.wcc.nrcs.usda.gov/data/water/wyfor/strm/misc/adj/

Tom Pagano also led the implementation of a daily statistical forecasting system for seasonal water supply outlooks, which will provide operational baselines for seasonal forecasts not made on the 1st of the month: ftp://ftp.wcc.nrcs.usda.gov/data/water/wcs/daily_forecast/

To improve the portability of UW-based models and forecasting, UW (Andy Wood) initiated a project to embed the VIC model within USDA/ARS’s Object Modeling System (OMS), a java-based modeling platform which will add a front/back end for the VIC model, as well as an interface for collaboration with interested parties. A stand-alone Trinity R. (CA) model was developed to serve as a prototype. This effort was suggested by Tom Pagano.

Andy Wood set up a website for data links for this testbed (that is linked to the main Hepex site):

For one of the HEPEX forecast basins, UW’s Ted Bohn has been exploring the value of using multiple land models for ensemble forecasting. A paper describing the research is in preparation.

Andy Wood has begun work to install a hydrologic ensemble forecasting capacity at NCEPEnvironmentalModelingCenter’s NLDAS core project. This project may lead to greater availability of ensemble hydrologic forecasts for the NLDAS domain (125W – 67W, 25N-53N).

3. Plans for the coming year

NWCChas been exploring the use of PRMS models for forecasting, and plans to calibrate and evaluate basins that may be contributed to the HEPEX testbed. Currently a number of other basins are running in real-time at NWCC with results available at:

ftp://ftp.wcc.nrcs.usda.gov/data/water/wcs/prms/.

UW’s Andy Wood has had two new grants funded that are expected to begin this summer or fall, focusing on snow assimilation and hydrologic forecasting. The first focuses on the assimilation of NRCS SNOTEL SWE observations into hydrologic models, and the second, a 3-year project, focuses on the assimilation of MODIS imagery and also on the use of simulated predictors in a statistical framework. Much of this research will be relevant tothe first objective of this testbed.

UW’s Andy Wood & Dennis Lettenmaier are directing research into the value of calibration for seasonal and shorter lead. The paper is expected to be finished in Fall 2006.

BC Hydro will recalibrate the Columbia River above Mica Dam basin in Winter/Spring 2006/07. The focus will be on developing a strategy to reduce simulation bias in general and the modeling error of ESP forecasts in particular. The use of automatic calibration in characterizing parameter uncertainty will be investigated.

4. List of publications and presentations

Publications

Weber, F., L. Perreault and V. Fortin, 2006: Measuring the Performance of Hydrological Forecasts for Hydropower Production at BC Hydro and Hydro-Quebec, Proceedings of the AMS.

Presentations

Wood, A.W., and D.P. Lettenmaier, Joint Meeting of the AGU, Baltimore, May, 2006, Using CPC long lead climate outlooks for ensemble streamflow forecasting.

Weber, F., L. Perreault and V. Fortin, 2006: Measuring the Performance of Hydrological Forecasts for Hydropower Production at BC Hydro and Hydro-Quebec, European Geosciences Union General Assembly, Vienna, Austria.

5. Links to related web sites, references, etc.

Pagano et al., 2004: Official NRCS water supply forecast evaluation paper:

ftp://ftp.wcc.nrcs.usda.gov/downloads/factpub/wsf/Pagano_et_al_JHM_2004.pdf

Pagano and Garen, 2003: Methodology for creating baseline forecasts for daily flow:

ftp://ftp.wcc.nrcs.usda.gov/downloads/factpub/wsf/Pagano_and_Garen_WSC_2003.pdf

A repository of links to test-bed data, publications, and announcements:

A brief description of the SWE data assimilation procedure used at BC Hydro: