INSIGHTS Declines in Leaf Litter Nitrogen Linked to Rising Temperatures in a Wet Tropical Forest Katherine Tully 1 and Deborah Lawrence Department of Environmental Sciences, University of Virginia, Clark Hall, PO Box 400123, Charlottesville, Virginia 22904, U.S.A. ABSTRACT In the tropics rainfall can vary by hundreds of millimeters from month to month, while mean temperatures fluctuate by only a few degrees. Nevertheless, during this 7-year study, we observed 35–52 percent declines in litter nitrogen concentrations in response to small increases in minimum temperature, with no response to the larger oscillations in rainfall. Abstract in Spanish is available at http://www.blackwell-synergy.com/loi/btp. Key words: Costa Rica; minimum temperature; rainfall; secondary forest. THE TROPICS ARE CURRENTLY UNDERGOING INTENSE LAND USE CHANGE, and as a result, secondary forests are increasing in abun- dance. Secondary forests cover an estimated 11 million hectares in Central America, of which 1.3 million are located in Costa Rica (FAO 2005). In light of global climate change, it is crucial to un- derstand how forest functions will respond to changing rainfall and temperature regimes. As nitrogen (N) is a major component of both chlorophyll and photosynthetic enzymes in leaf tissues, any changes in leaf N concentrations can affect photosynthetic rates (Murata 1969, Na ´tr 1975) with implications for regulating plant pro- ductivity and carbon uptake. This 7-yr study, explored how leaf litter nitrogen concentrations in secondary forests responded to rainfall and temperature variability. Rainfall in the wet tropics is seasonal and abundant. Although the intra-annual patterns in rainfall have been shown to drive changes in nutrient cycling in tropical dry forests (Murphy & Lugo 1986, Martinez-Yrizar & Sarukhan 1990, Read & Lawrence 2003), the dynamics and causes of litter nutrient variability have been little studied in tropical wet forests. We predicted that extended periods of high rainfall would lead to higher nitrogen concentrations in leaf litter deposited after the rains as a result of greater diffusion of nu- trients through the soil to the roots (Kobe et al. 2005, Wood et al. 2005). Alternatively, short periods of low rainfall could also lead to premature senescence (Wright & Cornejo 1990) and thus incom- plete nutrient resorption (Sizer et al. 2000) registering as high nitrogen content in litter shed following dry spells. Compared with rainfall, variation in temperature is subtle in the tropics, and minimum and maximum temperatures fluctuate far more throughout the day than the amplitude of seasonal change (Sanford et al. 1994). To our knowledge, no studies to date have focused on nutrient dynamics in tropical secondary forests in re- sponse to changes in temperature (but see Reich & Oleksyn 2004 for an analysis of global foliar nutrients in relation to temperature). Nevertheless, as tropical rain forests represent one of the warmest terrestrial ecosystems, they may be among the first to show negative responses to global changes in temperature (Saxe et al. 2001), and warrant an investigation of their response to temperature variabil- ity. We predicted that any change in litter N to temperature would be accompanied by a concomitant change in litter production (a proxy for leaf productivity). That is, leaf flushing corresponds to peaks in photosynthetically active radiation (Wright & van Schaik 1994) which are primarily driven by seasonal cloud cover variations (de Rocha et al. 2004). Cloud-free days in La Selva are weakly correlated with higher temperatures (Clark et al. 2003), and as temperatures increase, we expected a corresponding increase in plant productivity. As leaf flushing provides a sink for plant nutri- ents, essentially diluting the plant’s nutrient pool (Chapin et al. 2002), we expected litter N to decline following periods of elevated temperature. This study was conducted in secondary, tropical wet forests in and near La Selva Biological Station in Costa Rica (10126 0 N, 84100 0 W; Organization for Tropical Studies [OTS]; Holdridge 1947; all stands located within 14 km of one another). Permanent sample plots of 1 ha were established at four sites in 1997 (Chazdon et al. 2005). Before their protection for research, these stands were cleared by slash-and-burn, and were in pasture for 5–10 yr. In 2000, stands were 15–28 yr old (Chazdon et al. 2005; Table S1). They are located on highly weathered, volcanic soils that are pre- dominantly clayey with low bulk density, high organic matter con- tent, and high porosity (Parker 1994, Sollins et al. 1994). Mean annual rainfall from 1992 to 2007 was 4340 mm. Average monthly rainfall never fell o 150 mm, indicating that stands were probably not water-stressed during the study period. Yearly precipitation at La Selva is bimodal with 4 400 mm/mo in June–July and November–December. The period of least rainfall occurs between February and April with March being the driest month (Sanford et al. 1994). We analyzed litter collected between October 2000 and Octo- ber 2007. Fiberglass mesh traps (1 mm) were suspended 80 cm off Received 6 November 2009; revision accepted 21 May 2010. 1 Corresponding author; e-mail: ktully@virginia.edu BIOTROPICA 42(5): 526–530 2010 10.1111/j.1744-7429.2010.00693.x 526 r 2010 The Author(s) Journal compilation r 2010 by The Association for Tropical Biology and Conservation