Response of net ecosystem production of a coastal Douglas-fir forest to photosynthetically active radiation and saturation deficit

Three years of eddy covariance measurements of CO₂ flux above a coastal 55-year-old Douglas-fir forest were used to determine the individual and combined effects of diffuse photosynthetically active radiation (PAR), direct PAR and atmospheric saturation deficit (D) on net ecosystem production (NEP). A variable χ was defined as: χ = PAR_dif + k PAR_dir, where PAR_dif and PAR_dir were measurements of half-hourly diffuse and direct PAR, respectively, and k was a fraction increasing from 0 to 1. When k = 0, χ = PAR_dif and when k = 1, χ is total PAR. Correlating χ with the half-hourly NEP measurements revealed that (1) the PAR_dif generally accounted for significantly more variance in NEP than total PAR itself, (2) adding a certain fraction of PAR_dir (e.g., k = 0.2) always accounted for significantly more variance in NEP than when PAR_dif was used alone, and (3) the correlation coefficients between χ and NEP were generally significantly higher in low D conditions than in high D conditions. Results (1) and (2) were found to be independent of solar elevation angle, temperature, u_* (friction velocity), D, and total PAR, and thus suggested a direct effect of diffuse PAR on NEP. The values of k obtained in (2) indicated a significant scattering and penumbral effect of the direct solar beam. These effects are ignored in most canopy radiative models.

In addition, we divided the half-hourly NEP measurements into four combinations: cloudy high D, cloudy low D, sunny high D and sunny low D, where low and high D were defined as D < 1 kPa and D > 1 kPa, respectively. The response of NEP to total PAR was highest in cloudy low D conditions and lowest in sunny high D conditions. The responses of NEP to total PAR in the other two conditions were similar in magnitude and fell roughly midway between the highest and lowest NEP responses. This study showed that the response of NEP to total PAR in sunny conditions was linear, while in cloudy conditions it was parabolic. The frequently observed hyperbolic canopy photosynthetic light response was found to be the average of linear and parabolic responses over a period of several weeks that included sunny and cloudy days.

Results from this study provide a new perspective in understanding the impacts of volcanic eruptions (e.g., the effect of high PAR_dif), El Niño events (e.g., the effect of high D), and the frequent co-occurrence of these events on the global terrestrial carbon cycle.