Large-scale changes in climate may have many unexpected effects on ecosyste
ms, given the importance of climate as a control over almost all ecosystem
attributes and their many internal feedbacks. In particular, the interactio
ns among energy flux, plant dynamics, and soil carbon and nutrient cycling
are poorly known. In this study, we examined biotic controls over soil temp
erature and evapotranspiration (ET) in a climate change experiment in two p
eatlands, a bog and a fen. Bogs are isolated from groundwater inputs (i.e.,
ombrogenous) and are acidic, whereas fens receive groundwater inputs (i.e.
, minerogenous) and are more alkaline. They also have many associated diffe
rences in soil chemistry, nutrient availability, and plant communities.
We removed 27 intact peat monoliths each from a bog and a fen in northern M
innesota to construct a large mesocosm facility that allows for direct mani
pulation of climatic variables in a replicated experimental design. The tre
atment design was a fully crossed factorial with three infrared-loading tre
atments, three water table treatments, and two ecosystem types (bogs and fe
ns), with three replicates of all treatment combinations. Heating was achie
ved with overhead infrared lamps, so soil temperature was a dependent varia
ble.
Increased infrared loading caused mean monthly soil temperature to increase
by 1.6 degrees-4.1 degrees C at 15 cm depth during the growing season (May
-October). The soil temperature response was strongly seasonal, with much g
reater effects during the growing season than during the winter. Within the
growing season, heating caused greater increases in soil temperature in Ma
y-June in the bog plots and in May-June and September-October in the fen pl
ots. Occasionally during the winter, increased infrared loading cooled the
soil by up to 5.1 degrees C, probably due to melting of the insulating snow
cover. Fen plots were on average 0.8 degrees-1.0 degrees C warmer than bog
plots during the growing season, although bog plots had a steeper soil tem
perature gradient with depth. Water level had no effect on soil temperature
.
Annual ET was 12-23% greater in the bog plots than in the fen plots, at lea
st partially explaining the cooler soil temperatures in the bog plots due t
o latent heat of evaporation. Additionally, warmer, wetter conditions incre
ased ET. In the fen plots, ET was more sensitive to water table depth than
in the bog plots. Differences in the composition and seasonality of the pla
nt canopies accounted for much of the observed differences in ET between bo
g and fen mesocosms, and hence soil temperature. We present a conceptual mo
del of how above- and belowground ecosystem processes control energy fluxes
in northern peatlands and suggest that such controls represent a general p
henomenon that should be incorporated into climate change models that inclu
de biotic feedbacks.