Diel patterns of soil respiration in a moist subtropical forest: key drivers and future research needs

Moist tropical forests have the highest soil respiration rates (Rs) of any terrestrial ecosystem and account for approximately one third of the world's soil carbon (C) pool. Small increases in the magnitude of Rs in these ecosystems can result in high rates of soil C loss, with significant consequences for global climate change. Identifying the climatic controls of Rs in moist subtropical forests will improve our ability to predict how this large C flux will respond to climate change. Our objectives were (1) to determine whether Rs varies on diel timescales, (2) whether diel Rs patterns vary seasonally, and (3) identify biophysical drivers of this temporal variation. We measured hourly Rs in a secondary, moist subtropical forest in Puerto Rico for a 3-year period using an automated soil respiration system (LI-COR 8100/8150 with six chambers). Concomitant with Rs we measured hourly variation in several climatic drivers (air/soil temperature, soil moisture, relative humidity, and photosynthetically active radiation). Soil respiration showed significant diel variation, with the magnitude, amplitude, and shape of these curves varying throughout the year. Overall, diel Rs peaked in the late afternoon and reached a minimum in the early morning. Diel amplitudes ranged from 1 to 7 μmol CO2 m-2 s-1, with larger amplitudes occurring in warmer months that also have higher rates of Rs. In warmer months Rs exhibited a strong bimodal pattern, and a narrower diel range with a single peak in cooler and drier months. Diel Rs was positively correlated with soil temperature, but this relationship was non-linear during the day and linear at night (i.e., hysteresis). The bimodal pattern of Rs may be due to a mid-day depression of photosynthesis when humidity is low and air temperature is high, thereby reducing transport of photosynthate to the roots and decreasing rhizospheric respiration. The hysteresis between Rs and temperature suggests multiple controls on Rs on diel time-scales. Research that partitions Rs into its components could provide insight into their respective sensitivities to different climatic drivers, improving our capacity to understand the effects of climate change on the tropical forest C cycle.