Verification of Spot Fire Weather Forecasts
Principal Investigator(s):
  • John D. Horel
    University of Utah, Department of Meteorology
Co-Principal Investigator(s):
  • Thomas T. Lindley
    National Oceanic and Atmospheric Administration (NOAA)
  • Glen W. Sampson
    National Oceanic and Atmospheric Administration (NOAA)
  • John P. Snook
    US Forest Service
  • Jesse C. Pugh
    University of Utah
  • Sally P. Petersen
    University of Utah
Completion Date: August 25, 2015

Cataloging Information

  • fire weather
  • forecasting
  • NDFD - National Weather Service National Digital Forecast Database
JFSP Project Number(s):
Record Maintained By:
Record Last Modified: December 13, 2016
FRAMES Record Number: 21097


The proposed work will assess the degree of improvement provided by spot and incident fire weather forecasts as compared to National Digital Forecast Database (NDFD) forecasts and provide a methodology to verify fire weather forecasts nationally. The expected benefits from this project include: 1) improved forecast accuracy of spot forecasts for both prescribed fire and wildfire events nationally resulting from the feedback obtained by objective metrics as well as user comments; 2) recommended path for implementation of verification tools within the NWS Performance Branch; 3) increased numbers of potential burn days, due to dependable verification statistics, to better enable the fire managers to meet their target goals and objectives. We propose to evaluate critical elements of spot forecasts: surface weather (e.g., temperature, relative humidity, and wind) and boundary layer parameters (e.g., mixing height, Haines Index, clearing index, smoke dispersal). We will examine spot forecasts (~20,000 per year nationally) over the period from 2011 through the completion of the project compared to other baseline forecasts such as persistence and climatology. While assessing the quality of all spot forecasts issued is important, evaluating forecasts issued at critical times in critical locations is particularly relevant for the safety of fire personnel and management of natural resources. Our work is intended to allow the flexibility to assess and evaluate fire weather forecasts as they pertain to local situations recognizing that the spot forecast elements requested by users from NWS Forecast Offices varies substantively around the nation. Spot forecasts for surface weather elements will be compared to NDFD forecasts and validated against in situ and nearby NWS and RAWS observations, and 2.5 km and 5 km RTMA gridded analyses. Boundary layer elements will be compared to output from the NCEP Rapid Refresh hourly analyses. A critical aspect of our study is to evaluate the uncertainty in the verifying data sets in order to define appropriate bounds for defining skillful vs. unskillful forecasts. We propose to validate the forecasts from three perspectives: (1) strategic verification- to provide flexible tools for evaluating spot forecasts locally, regionally, and nationally; (2) forecast improvement- to provide NWS forecasters with the capability to examine prior forecasts within the framework of their operational forecast preparation process; and (3) forecast usage- identifying from users the impact of the forecasts on fire and incident operations. Project deliverables include the results of the spot forecast verification analysis that will help to understand characteristics of spot forecasts as a function of location, season, and weather situation. In addition, we will foster implementation of a forecast methodology appropriate for spot forecasts in the Statistics-On-Demand operational system managed by the NWS Performance Branch.