|Scott, Russell - Russ|
Submitted to: Journal of Geophysical Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/28/2009
Publication Date: 11/4/2009
Citation: Scott, R.L., Jenerette, G., Potts, D., Huxman, T. 2009. Effects of seasonal drought on net carbon dioxide exchange. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114: 1-13. G04004. Interpretive Summary: The increase in the amount of woody plants (shrubs and trees) into grassland ecosystems is one of the most pervasive changes in land cover in the southwestern US and around the world. It is important to understand how this change in vegetation will affect water and nutrient cycling of ecosystems in order to predict the outcomes of this change on society. We measured energy, water and carbon dioxide exchange between a woody-plant-encroached grassland and the atmosphere over a four-year period and determined how the amount of precipitation controlled these exchanges. In contrast to the current paradigm that woody plant encroachment might result in more ecosystem carbon sequestration and to the many recent results showing that various semiarid ecosystems were a sink for carbon dioxide, we found that this ecosystem was a source, which was likely a consequence of the decade-long drought that was on-going over the study period. These results highlight a complex relationship between vegetation change and climatic variation in precipitation that likely influences the carbon sequestration potential of water-limited landscapes.
Technical Abstract: The encroachment of woody plants into historical semiarid grasslands has important ecohydrological and socioeconomic consequences. In this paper, we document the biosphere-atmosphere exchange of water and carbon dioxide that occurred from 2004 through 2007 over a semiarid, warm-season savanna in southern Arizona, USA. Over the last 100 years, this historic grassland has converted to a native mesquite (Prosopis veluntina Woot.) savanna. We used eddy covariance to measure material and energy fluxes along with standard meteorological and soil moisture measurements to better understand the controls on these exchanges. Each year of the study had below normal precipitation (P), but monsoon (July-September) P was both above and below average, while winter/spring (December-March) P was always less than average. Each year, cumulative evapotranspiration (ET) prior to the beginning of the rainy season (~January through June) exceeded P with the situation usually reversing itself by the end of the year. Over the four-year period, total ET (1234 mm) nearly matched P (1239 mm), indicating that there was little surface runoff and no deep drainage. ET was significantly correlated with shallow (0-10 cm) soil moisture year-round, but was also correlated with soil moisture down to at least 1 m during the middle to late growing season. Diurnal patterns of net ecosystem exchange of CO2 (NEE) revealed the dominant influence of the July-September rainy season on the CO2 exchange, although a smaller, shorter duration growing season in the months of March-May was apparent in years with relatively higher winter/spring rainfall. Annual photosynthesis (GEP) was strongly correlated with monsoon precipitation (p < 0.05), but respiration and NEE were not significantly correlated with either monsoon or annual precipitation. Ecosystem respiration, facilitated by abundant shallow soil moisture during the monsoon, was considerably higher in 2006 following a very dry winter and spring. During these four drought years, the ecosystem appeared to be net source of CO2 to the atmosphere, ranging from 14 – 95 g C m-2 yr-1 with the strength of the source increasing with decreasing precipitation. In 2005, the year with the closest to average rainfall, the ecosystem was nearly carbon neutral. We suggest that below-average rainfall conditions during the period of study were especially detrimental to mesquite growth. The growing season of the trees may span both spring and summer periods, but because of the reduced winter/spring input, the spring growing season was significantly truncated or completely absent. Thus, the ecosystem was likely “burning off” much of the carbon that may have been sequestered during the 15-20 year period of wetter-than-average rainfall that preceded the present drought. These results highlight a complex relationship between vegetation change and decadal variation in precipitation that likely influences the source/sink dynamics of water-limited landscapes.