Long-term productivity in a tropical dry forest: Topographic and climatic variability

Twenty-one years of litterfall data collected at the Chamela tropical dry forest (Pacific Coast of Jalisco, Mexico) revealed strong interannual differences in productivity along an elevation gradient (60–160 masl) within a watershed (16 ha). Annual precipitation during the study period varied from 366 to 1261 mm, with unusually heavy rains (>100 mm) in the dry season during ENSO years. Annual litterfall was significantly and consistently higher in the bottom of the watershed (range 350-618 g m^-2), where potential interception of solar radiation is lower, and soils are deeper and higher in water content. Litterfall in the middle and upper positions were lower and with a narrower range (252-453 g m^-2). Litterfall and current year rainfall were not significantly correlated in any plot, but correlation with previous year rainfall was significant only in the bottom plot. This highlights the influence of topography in soil water availability and productivity. The significant (P<0.02) correlation between annual litterfall and dry season (Dec-May) rainfall showed that years of unexpected rains during the dry months may trigger productivity up by a factor of 1.25, compared to the long-term average on each landscape position within the watershed. Temporal changes in litterfall followed the seasonal pattern of rainfall, but seasonality of litterfall (coefficient of variation of monthly litterfall) was positive and significantly correlated with mean temperature/mean rainfall ratio. Long-term data are crucial to analyze forest functional processes, where interannual and decadal scale variation in climate is great and plant performance is strongly constrained by the physical environment.