A study of interpolation methods for forest fire danger rating in Canada

Authors:
Publisher website
Google Scholar
Abstract
Canadian fire control agencies use either simple interpolation methods or none at all in estimating fire danger between weather stations. We compare several methods of interpolation and use the fire weather index in the North Central Region of Ontario as a case study. Our work shows that the second order least square polynomial, the smoothed cubic spline, and the weighted interpolations had the lowest residual sum of squares in our verification scheme. These methods fit the observed data at both high and low fire weather index values. The highly variable nature of the spatial distribution of summer precipitation amount is the biggest problem in interpolating between stations. This factor leads to highly variable fire weather index fields that are the most difficult to interpolate. The use of radar and (or) satellite data could help resolve precipitation patterns with greater precision. These interpolation methods could easily be implemented by fire control agencies to gain a better understanding of fire danger in the region.