Present-day global predictions were 645 PgC in vegetation, 1190 PgC in soils, a mean carbon residence time of 40 years, NPP 47 PgC y(-1) and NEP (the terrestrial sink) about 1 PgC y(-1), distributed at both high and tropical latitudes. With IS92a emissions, the high latitude sink increased to the year 2100, as forest NPP accelerated and forest vegetation carbon stocks increased. The tropics became a source of CO2 as forest dieback occurred in relatively hot, dry areas in 2060-2080. High VPDs and temperatures reduced NPP in tropical forests, primarily by reducing stomatal conductance and increasing maintenance respiration. Global NEP peaked at 3-4 PgC y(-1) in 2020-2050 and then decreased abruptly to near zero by 2100 as the tropical source offset the high-latitude sink. The pattern of change in NEP was similar with CO2 stabilization at 750 ppm, but was delayed by about 100 years and with a less abrupt collapse in global NEP. CO2 stabilization at 550 ppm prevented sustained tropical forest dieback and enabled recovery to occur in favourable years, while maintaining a similar time course of global NEP as occurred with 750 ppm stabilization. By lessening dieback, stabilization increased the fraction of carbon emissions taken up by the land. Comparable studies and other evidence are discussed: climate-induced tropical forest dieback is considered a plausible risk of following an unmitigated emissions scenario.
Google Scholar Search: links, citations and journal (if available)