Mapping the reduction in carbon uptake in subarctic birch forests due to insect outbreaks.

It is projected that forest disturbances, such as insect outbreaks, will have an increasingly negative impact on forests with a warmer climate. These disturbance events can have a substantial impact on forests' ability to absorb atmospheric CO₂, and may even turn forests from carbon sinks into carbon sources; hence, it is important to develop methods to both monitor forest disturbances and to quantify the impact of these disturbance events on the carbon balance. In this study we present a method to monitor insect induced defoliation in a subarctic birch forest in northern Sweden, and to quantify the impact of these outbreaks on gross primary productivity (GPP). Since frequent cloud cover in the study area requires data with high temporal resolution and limits the use of finer spatial resolution sensors such as Landsat, defoliation was mapped with remote sensing data from the MODIS sensor with 250 × 250 m spatial resolution. The impact on GPP was estimated with a light use efficiency (LUE) model that was calibrated with GPP data obtained from eddy covariance (EC) measurements for years with undisturbed birch forest and for years with insect induced defoliation. Two methods were applied to estimate the impact on GPP: (1) A GPP reduction factor derived from EC-measured GPP was applied to estimate GPP loss, and (2) a LUE model was run both for undisturbed and defoliated forest and the differences in modelled GPP were derived. In the study area of 100 km2 the results showed a total decrease in carbon uptake over three outbreak years 2004, 2012, and 2013 of 44.6 ± 13 Gg C, which is of the same magnitude as the estimated annual GPP of 41.1 ± 12 Gg C for a year without disturbance. In the most severe outbreak year (2012), 76% of the birch forests were defoliated and annual GPP was merely 50% of GPP for years without disturbances. The study has generated valuable data that improve previous studies on impact estimates and demonstrates a potential for mapping insect disturbance impact over extended areas. [ABSTRACT FROM AUTHOR]