The models used for projecting long-term ecosystem response to climate change mostly parameterized by short-term data. However, ecosystem responses to global change are strongly regulated by long-term processes, such as species replacement in communities. So, the short term ecosystem response to climate change may be different from long term. In order to comparing the short vs long term response of ecosystem carbon cycling to climate warming, we conducted two manipulative warming experiments in a tallgrass prairie in the Great Plains. The long-term experiment used a paired nest design with warming as the main factor nested by clipping factor. The treatments were conducted since September 1999. The short-term experiment used a completely random design with full interaction of two warming treatments (ambient and warming) and three precipitin treatments (ambient, double, half precipitation), which were treated since July 2009. Net ecosystem carbon exchange (NEE), gross primary productivity (GPP), and ecosystem respiration (ER) were measured by Licor infrared gas analyzer connected to a static chamber.

Results/Conclusions: The results showed that warming decreased GPP, ER, and NEE in the short term but increased them in the long term. The different short vs long term response may primarily due to the changes in species composition. The short term response of carbon cycling to warming was a function of the initial C3 dominance structure, whereas the long-term response reflected the shifts of species to C4 dominance. Although we only have seven months data up to now, the results do suggest that initial response of ecosystem carbon cycling can not represent the long term response. Ecological models should consider long term vegetation dynamics to more accurately predict climate change effect.