Forest Investment Account (FIA) - Forest Science Program
FIA Project Y081107

    Improvement of a mechanistic risk model for estimating windthrow losses
Project lead: Mitchell, Steve (University of British Columbia)
Author: Mitchell, Stephen J.
Subject: Forest Investment Account (FIA), British Columbia
Series: Forest Investment Account (FIA) - Forest Science Program
Wind damages trees in unmanaged and managed stands and both managers and modellers need better tools to predict windthrow losses. Since 2005 we have been working with the BCMOF Decision Support group and the UK Forestry Commission to integrate the UK Forestry Commission’s mechanistic windthrow risk model ForestGALES with the growth and yield model TASS/TIPSY (FSP Project Y072169). ForestGALES is a non-spatial model that predicts the critical wind speed for tree failure and the probability of this critical wind speed in even-aged stands. We have updated ForestGALES by adding mechanical properties of BC conifers from our recent winching and wind tunnel studies (hereafter referred to as ForestGALES_BC). We have also improved the equations for tree stem and crown representation using values derived from TASS and from our field measurements. We have replaced the regional/topographic windiness score (DAMS) used in the UK with gridded mean annual wind speed data from BC Hydro’s Wind Resource map and local topographic exposure scores to estimate local probability of a given wind speed and direction. We have also developed an interface, WINDFIRM, which characterizes within-opening wind exposure due to fetch and boundary orientation. We have successfully used WINDFIRM to run stand-level empirical windthrow models (Lanquaye and Mitchell 2005, Scott and Mitchell 2005) and display relative windthrow risk for cutblock edge segments. As part of the ForestGALES/TASS/TPSY project, we have converted the ForestGALES source code from Pascal to open-source Python, and WINDFIRM from an ArcView extension in Avenue scripts to Python. We have updated the ForestGALES code by adding critical turning moment and drag equations obtained from our winching and wind tunnel studies (e.g. Rudnicki et al. 2004, Byrne and Mitchell, in press). We are also updating dendrometrics (tree allometry) using equations and tree outputs from TASS (Mitchell 1975). WINDFIRM was designed to spatially represent cutblock edge conditions by assembling data from layers such as forest cover attributes, above-canopy wind speed, topographic exposure for cutblock edge segments and passing this data to logistic regression models for calculation of windthrow probability. We have converted WINDFIRM into an interface that processes and assembles data for gridpoints at regular intervals across an area, combines this with tree list data and passes the data to ForestGALES. Through our collaboration with Dr. Roland Stull in FSP Project Y062276 we archived and summarized one year of hourly weather forecasts. While Y062276 ended in March 2006, Dr. Stull continued to archive data and now has a second full year. This dataset provides the opportunity to further improve the characterization of wind directionality and speed at a resolution of 2 km across the southern half of the province. In the course of the project to integrate ForestGALES with TASS/TIPSY, we have identified a number of limitations in ForestGALES. The purpose of the proposed research is to add greater versatility to the model so that it better reflects conditions in BC stands and takes advantage of new data sources. ForestGALES was designed for very homogeneous, single species UK plantations and assumes that the mean tree represents all trees in the stand. Consequently, the entire stand either survives or fails at a given wind speed. In reality, even in uniform stands, trees form various crown classes and differ in windfirmness. Loss of neighbouring trees increases wind loading on remaining trees, and damage propagates until the remaining trees are stable enough to survive the incremental exposure. At the present time, ForestGALES can only simulate continuous openings or uniform thinnings. Furthermore, ForestGALES treats wind within canopy gaps as non-directional. BC is more complex topographically than Britain and recent developments in numerical weather prediction modeling at UBC give us the opportunity to incorporate better, more localized wind data for strong wind events. This project will build on work completed in previous research and complement the ForestGALES/TASS/TIPSY integration project to produce a mechanistic windthrow model with the versatility to accommodate non-uniform young mature and mature stands using tree-lists, simulate damage propagation, and incorporate wind directionality and directional fetch (exposure within openings). These represent major advances in model sophistication. We will be better able to represent the tree, stand, soil, topographic and wind conditions in the silviculture systems trials for which we have data and this will simplify model validation. More reliable predictions in uniform and non-uniform stands will result in more realistic simulations when ForestGALES_BC is released with TASS version 3, for both complex harvest and no-harvest scenarios.


Final Report (1.7Mb)

To view PDF documents you need Adobe Acrobat Reader, available free from the Adobe Web Site.

Updated August 16, 2010 

Search for other  FIA reports or other Ministry of Forests and Range publications.

Please direct questions or comments regarding publications to