Figure 1. Monthly estimates of carbon exchange by the old-growth hemlock stand during day and night periods from November 2000 through October 2001. Net carbon storage is the sum of daytime carbon storage and nighttime carbon loss.

Figure 2. Carbon exchange measured by eddy covariance over 30-minute intervals with wind blowing from the hemlock stand toward the sensor (SW wind). Note that in this figure carbon storage is plotted as a negative number, the reverse of Figure 1. Large gaps in data are due to computer failure (December, January and early August), and removal of the flux system for testing against the other Harvard Forest eddy covariance system (late August and early September.) Shorter data gaps are primarily due to periods when wind was not from the SW. Heavy solid lines in both the daytime (upper) and nighttime (lower) panels show model estimates of ecosystem carbon exchange derived from eddy covariance data. The models predicted C exchange by multiple regression analysis of environmental variables significantly related to measured carbon exchange. These variables included photosynthetically active radiation, air temperature, daily minimum air temperature, and soil temperature. Lightly broken lines (obscured by data in many areas) are estimates of soil respiration from chamber measurements and an exponential temperature response model.

Figure 3. Response of hemlock forest carbon storage to photosynthetically active radiation (PAR) after nights with various minimum temperatures (Tm) in early spring 2001. Regression lines show best-fit regressions of CO2 flux to the log of PAR, and r2 values refer to these regressions. A lack of response to PAR indicates suppressed photosynthesis.