We use the eddy covariance method from Sep 2008 to December 2010 to measure CO 2 fluxes between the atmosphere and the regenerated dry dipterocarp forest in western Thailand. The dry dipterocarp forest ecosystem is typically a characteristic of the area where a clear dry period of 3-4 month and precipitation amount of 900-1200 mm yr-1 is found. This was occurring at our flux site during the dry period of the end of 2008 to the beginning of 2009. However, in 2010 the unusual long dry period (5 months) during December 2009-April 2010, and the late coming and long lasting raining period occurred over several areas of Thailand including our flux tower site. We hypothesized that such unusual long dry period could affect the dynamics of CO 2 exchange. In this paper we report the results of CO 2 exchange during such period and compare them to the data collected during the normal years of 2008-2009. The study site is located within the 88.9 ha-dry dipterocarp forest inside the campus of King Mongkut's University of Technology Thonburi (KMUTT), Ratchaburi Province in western Thailand (13° 35' 13.3'' N, 99° 30' 3.9'' E). We found that GPP and RE differed in their responses to the increase in soil moisture. This is especially obvious at the end of dry season when rain comes. During the transition period from dry to wet seasons, RE increases more rapidly than GPP. As a result, during such particular period the NEP significantly decreases due to the increasing contribution of RE. This is one of the important characteristics of the dry dipterocarp forest, that NEP is lowest at the onset of raining season. On the other hand, we do not find the different response of GPP and RE during the transition from wet to dry period; both are decreased at similar rate. During the dry period when forest shades their leaves, the carbon exchange process is dominated by RE. We also found that the unusual long dry period of 2009-2010 do not have any significant effects on the annual NEP (12.06 and 12.26 ton C ha-1 year-1 for 2009 and 2010, respectively). This is because the increase in RE during early raining season is compensated by the high GPP during the late raining season. Thus, this dry dipterocarp forest seems to effectively adapt to variations in climatic variables such as long dry period. We currently investigate the energy balance and water use efficiency, which would help us understand more about the various processes and adaptive mechanisms of this forest ecosystem to future extreme climate events. This is because the increase in RE during early raining season is compensated by the high GPP during the late raining season. The timing adjustment of early/late budding and shedding seems to be very important that makes this dry dipterocarp forest well adapted to short-term variations in extreme climate phenomena. ABSTRACT