| description abstract | Fire weather stations used to be located only where observers were available. Now, remote automatic weather stations can sample weather without manual assistance, virtually anywhere, Locating them, however, remains a problem. An objective method was developed for designing a fire weather network. The design process begins with the identification of the area to be sampled, the relevant meteorological variables, target meteorological fields, and interpolation (or other analysis) method to be used. Using a least-squares criterion, optimum networks for a varying number of stations are computed with a constrained optimization algorithm. The optimum error sum of squares, thus obtained by interpolation, is used to help choose an optimum network. Simulated fire rate-of-spread fields were analyzed with optimum networks varying in an from one to seven stations. Preliminary results were obtained for the design of a fire weather network for southern California. | |
