Variation in carbon isotope composition among years in the riparian tree Populus fremontii

Water availability is an important factor limiting the productivity of desert plants but little is known about the impact of water-limiting conditions on the physiology of plants in mesic environments. Riparian ecosystems of the western US receive significantly more water than the surrounding desert environments but experience dramatic interannual fluctuations in water availability because both stream flow and precipitation are highly variable over time. This variability results in different growing conditions each year which may influence the physiology of riparian species such as Fremont cottonwood (Populus fremontii), the dominant, native canopy tree species in lowland southwestern US river systems. We wished to determine if the physiology of this species varies among years, what climatic parameters are related to any observed physiological variation and if individuals within a P. fremontii population differ in their physiological response to variation through time. We collected tree ring cores from a central New Mexico cottonwood population and analyzed carbon isotope composition (δ¹³C) in each year from 1981 to 1995. We used δ¹³C analysis in this study because it allowed us to obtain multi-year estimates of physiological activity. During these years, mean stream flow at our study site ranged over two orders of magnitude from 0.82 to 80.94 m³ s^-1, precipitation ranged fourfold from 49 to 215 ccmm and mean temperature ranged from 20.5 to 22.6°C during the growing season. δ¹³C varied from a low of -26.7‰ in 1984 to a high of -24.7‰ in 1981. Low δ¹³C values were associated with years in which stream flow and/or precipitation were high and temperature was low. The opposite was true of years with high δ¹³C values. We observed a strong linear relationship between δ¹³C and stream flow during years when stream flow was <25 m³ s^-1 but no significant relationship between these variables when stream flow was >25 m³ s^-1. Additionally, there was a linear relationship between δ¹³C and precipitation during years when stream flow was <25 m³ s^-1 but not in years when stream flow was >25 m³ s^-1. These data suggest that above a threshold of total stream flow, increased flow does not influence physiology. Below this threshold, precipitation can be an important water source. The ten individuals within our study population varied significantly in mean δ¹³C values but responded to interannual variation in a similar manner (i.e., all individuals had low δ¹³C values when water was abundant). These results suggest that precipitation as well as stream flow are important factors influencing the physiology of this riparian tree.