@article {bnh-8191, title = {Wind speed Reduction Factors (WRFs): utilities for WRF assessment and communication - Black Summer final report}, number = {688}, year = {2021}, month = {08/2021}, institution = {Bushfire and Natural Hazards CRC}, address = {MELBOURNE}, abstract = {

This project commenced in March 2021 as a Bushfire and Natural Hazards CRC Black Summer funded initiative between Queensland Fire and Emergency Services and the School of Earth and Environmental Sciences at the University of Queensland. The purpose of the project was to undertake an evaluation of WRFs used in Australia (including during Black Summer) to quantify the reduction of open space wind speed by Australian fuel types.

Wind is a key driver of fire behaviour and can be highly variable and difficult to predict, particularly within the lowest 1-2km of the atmosphere where it interacts in complex ways with topography and vegetation. Operational fire spread modelling quantifies the impact of vegetation or fuel type on wind speed using Wind speed Reduction Factors (WRFs) or Wind Adjustment Factors (WAFs). Specifically, these factors quantify the impact of vegetation on reducing the speed of the open space prevailing wind. WRF is typically the ratio of 10m open wind speed to 2m wind speed, whereas WAF is the ratio of {\textquoteleft}midflame{\textquoteright} wind speed to 20 ft open wind speed.

To date, single or static WRFs have been assigned to 62 Queensland Broad Vegetation Groups (BVGs) for use within the operational fire simulation application PHOENIX Rapidfire. These WRFs have been derived via approximation. Specifically, the 10m open wind speed has been approximated by the closest Bureau of Meteorology (BOM) Automatic Weather Station (AWS) to the field site, and the 2m near-surface wind speed has been approximated via the use of a handheld anemometer raised to eye-level at a suitable location within the field site.

The key research aims were to:

The review found that the use of approximated static WRFs has caused minor to significant error accumulation in the fire spread model outputs produced by fire simulation applications, including PHOENIX Rapidfire.

To reduce error in fire spread modelling, the review concluded that the development of dynamic WRF modelling capabilities should be a priority. These dynamic WRFs should respond to key wind, fuel, fire and topography parameters that change over time and space.

However, a dynamic WRF model should not require such high levels of computation so as to delay real-time fire spread modelling outputs. At its simplest, a dynamic WRF model is a discrete, empirically derived WRF profile, illustrating the change in WRF at specific heights measured within a fuel type in the field. A more advanced dynamic WRF model might be a mathematical model for which wind, fuel, fire and topography parameters act as inputs and a mathematically idealised continuous WRF profile is the output. This model should be validated by empirical data. Overall, each fuel type should have its own WRF profile. The end goal should be to replace all static WRFs with dynamic WRF profiles in fire spread models.

A WRF test site was established in the priority fuel type {\textquoteleft}moist to dry eucalypt woodland on coastal lowlands and ranges{\textquoteright} at the Queensland University of Technology (QUT) Samford Ecological Research Facility (SERF), located on the outskirts of Samford Valley in Southeast Queensland. Installed at the site is a 15m instrumented tower using 3D sonic anemometers to record mean 3D wind speed, vertical wind direction and sonic air temperature. A discrete WRF profile was derived by taking the ratio of the average 10m open wind speed measured by the nearest BOM AWS in Brisbane and the average wind speed measured at heights of 2.5m, 4.5m, 10.5m and 15.5m. This WRF profile is preliminary as it is based on 23 hours of data collected outside the southeast Queensland fire season (August {\textendash} December).

Preliminary investigations of relationships between variables related to WRF were also conducted. The overall wind profile was compared to the Plant Area Density (PAD) profile of the vegetation obtained via terrestrial LiDAR (Light Detection and Ranging). A weak to moderate relationship was identified (R2 = 0.22) between mean wind speed and PAD, which may be due to the calm conditions experienced over the short length of the data collection period. Additionally, the day-time and night-time subcanopy temperature profiles were compared. The day-time profile was found to be slightly more constant with height, which may indicate that the subcanopy environment is more mixed and turbulent throughout the day. This result was supported by increased measurements of vertical mixing throughout the day. Nevertheless, data collection over a longer period under more varied conditions is recommended to investigate these relationships further.

The WRF test site at the QUT SERF has provided a preliminary insight into the relationships between vegetation and meteorology in the Australian context, which is essential for the development of empirically based dynamic WRF profiles for all fuel types. The methodology used is transferable and will be applied to other sites containing other priority fuel types. Anemometer measurements and LiDAR scans may then be used as key datasets for underpinning and validating the development of advanced dynamic WRF modelling capabilities in the next generation of fire spread models. Until this capability is developed, the new quick-reference WRF profile assessment resource developed by this project will enable FBANs near the fire ground to quickly identify the WRF values most relevant to the ensuing fire spread. These values may then be communicated to fire spread modellers.

}, keywords = {assessment, black summer, communication, factors, reduction, speed, wind}, issn = {688}, author = {Hamish McGowan and Katherine Rosenthal and Raymond Bott and John Myles} }