Meteorological Network for Wild Fire Prevention Program
Wild Fire risk can be modeled for prevention and management of resources if good meteorological data is available. There are several tools developed for wild fire risk assessment that use temperature, relative humidity, radiation, wind and specific measurements like fuel temperature, as input for their modeling. It is well know how a precise measurement of these variables is necessary so errors are not amplified by the calculations giving wrong assessments.
Accuracy depends on calibrated excellent quality equipment but also on representative measurements. For these reasons not only good sensors are needed, also excellent locations are necessary. The measuring strategy of a wild fire prevention network does not necessarly be the same than a generic meteorological or climate network. This affects the location of sites but also aspects like temporal availability of data or intensity of the maintenance programs. For some civil protection applications, it seems a requirement to have an owned meteorological network.
Having an owned network requires more resources than using external sources of data but gives absolute control on its reliability and operation procedures.
The objective of this network is to obtain accurate, reliable, uninterrupted and almost online measurements of :
- wind velocity and direction
- 2 m height air temperature and relative humidity
- solar global radiation
- fuel temperature and humidity (dead wood)
All this data should be every hour in a central server in order to be used by assessment tools.
The equipment used for this network was first quality sensors and logging systems. Campbell Scientific CR1000 with TCP/IP communications and compact flash extension was used. GPRS communications was used for data reporting and an over dimensioned solar power unit was installed to prevent power losses. Hourly communications are responsible for mostly of the power consumption. Protecting fences were installed to prevent vandalism, since the locations are remote and relatively exposed to public.
For communications we used a GPRS router as gateway with TCP/IP connection in combination with Campbell NL115. This behaves as a web server which is very useful for remote checking of the performance even with a smart phone. FTP protocol is used for data uploading to a central server.
Results from this network have been almost perfect with not important failures. Mostly of the failures were due to interruptions of the GPRS service. Fortunately vandalism seems not to be a problem, and the integration with the assessment and prediction tools went smoothly thanks to the easy integration of the ASCII data files on any integrated system.