On the seasonality of isoprene emissions from a mixed temperate forest

Measurements of isoprene concentration and flux were made at a mixed decidu ous forest in southern Canada during 1995 to characterize diel and seasonal emissions and thus deduce annual inventories. Isoprene inventories are nec essary for inputs to modeling systems to study atmospheric chemistry and ca rbon budgets. Despite adequate environmental conditions to promote emission s, the onset of isoprene emission occurred two weeks after full leaf expans ion, and two additional weeks were required for plants to emit isoprene at the maximum capacity. Such maximum isoprene emission was measured during Ju ly when canopy isoprene fluxes reached 40-60 nmol (isoprene).m(-2) (ground area).s(-1). Isoprene emission precipitously declined in concert with autum nal leaf senescence, with fluxes reaching the detection limit before the fo rest became leafless, In addition to plant development controls on emission s, temperatures below 10 degrees C strongly modulated isoprene emission. Af ter plants were exposed to low temperatures, isoprene emission remained sup pressed and did not correspondingly increase in the manner that temperature is known to influence isoprene biosynthesis. Using a one-dimensional model to vertically adjust temperature and visible solar radiation with depth in the canopy, coupled with a seasonally adjusted emission rate, we estimated that the forest produced 71 mmol isoprene/m(2) during 1995. For a deciduou s forest with final leaf-area index of 4.1 and active isoprene biomass of 7 5 g (dry mass)/m(2), on average such isoprene source accounted for 2% of th e carbon fixed through photosynthesis. The percentage of carbon entering th e atmosphere in the form of isoprene became as high as 10% during warm (>30 degrees C) and dry conditions. The data set reported here demonstrates tha t constant emission rates are inadequate to characterize emission rates for the entire growing season. Improved isoprene-emission inventories can be a chieved if emission factors are seasonally adjusted. In this study we adopt ed a method to express the emission rates as a function of degree days.