Talk given at the Madrean Conference in Tucson AZ.
Co-authored with Stephen Yool and Donald Falk.
Measuring ecosystem phenology via remote sensing enables prediction of risk factors related to disturbance and overall ecosystem health. For example, estimating the (1) risk and potential severity of future wildfires, (2) likelihood of an insect or disease outbreak, or (3) severity of drought morbidity and mortality, are each informed variously by antecedent weather conditions and the physical characteristics of the organisms phenotypes, as measured from earth observing systems (EOS). Fuel Moisture Stress Index (FMSI) was first proposed by Yool (2001), as a means of measuring the variation in live fuel moisture and thus its relative rigor or stress. We provide an empirical basis for FMSI, in context, with EOS data from the past 40 years. FMSI is derived using the formula for the standard score, or z-score. It is computed using the pixel’s NDVI recorded for a given time step/period in a given year, and the NDVI mean and NDVI standard deviation for the same time step/period across the entire time series. Tracking the FMSI for each year for a given pixel’s ‘phenoperiod’ roughly yields, with some statistical rigor, interannual trends in (as we like to say) ‘climate forcing’ and are amenable to inferential analyses. Prediction of potential fire behavior from remotely sensed imagery is an essential task for fire managers. FMSI could be used to assess the level of local drying in live vegetation which contributes to the ERC and potential flammability of the landscape in near real time. Utilizing a publicly available cloud computing service, Google Earth Engine, and a version control system, Github, users can repeat our analyses or calculate FMSI anywhere else globally within the time period of the reference EOS satellite platforms at no cost.
Link to code hosted on Github