Impact of sky conditions on net ecosystem productivity of a “floating blanket” wetland in southwest China
Abstract. Based on eddy covariance (EC) measurements from 2016 to 2020, the impact of sky conditions on net ecosystem productivity (NEP) over Beihai wetland was examined. Sky conditions were classified into sunny, cloudy and overcast skies. On half-hourly timescale, the daytime NEP responds to the changing total photosynthetically active radiation (PARt) more efficiently under cloudy and overcast conditions than sunny conditions across seasons. Compared with sunny conditions, the apparent quantum yield (α) under overcast (cloudy) conditions increased 342.9 % (271.4 %) in spring, 17.6 % (20.6 %) in summer, 280.0 % (230.0 %) in autumn and 125.0 % (25.0 %) in winter, respectively. Unlike the patterns of half-hourly NEP, the daily NEP was significantly lower under overcast conditions than that under cloudy and sunny conditions. And the daily NEP peaked under cloudy skies when the clearness index (CI) fluctuated around 0.3–0.6. Additionally, the ecosystem light use efficiency (LUE) and water use efficiency (WUE) also changed with the variations in sky conditions. The daily LUE and WUE reached their maximum values under overcast (CI: 0–0.2) and cloudy conditions (CI: 0.2–0.4), respectively. NEP was mainly controlled by the diffuse photosynthetically active radiation (PARd) and air temperature (Ta), and the direct photosynthetically active radiation (PARb) had a secondary effect on NEP from half-hourly to monthly timescales. Path analysis revealed that PARb and Ta were the main controls affecting NEP under sunny conditions. While under cloudy and overcast conditions, PARd was the most responsible for the variations of NEP.
Yamei Shao et al.
Yamei Shao et al.
Yamei Shao et al.
Viewed (geographical distribution)
The manuscript presents an empirical analysis of the relationship between cloudiness and NEP. From the upstart I see a serious problem in the approach. Specifically, with regards to the variables they choose to analyze.
First, why look at effects on NEP? Radiation (par) has direct effects on photosynthesis and thus, GPP, and is not expected to affect Respiration. Any effect on respiration is indirect, either through increased sugar transport to the roots, or through increase temperature. Furthermore, in the analysis of WUE, why use ET and not transpiration? especially in a flooded wetland, the direct evaporation from the water surface and saturate soil related to radiation in a very different way that transpiration and has nothing to do with carbon uptake so it should not be included in WUE.
Second, the key driver of this analysis, cloudiness, was not observed, instead it is derived indirectly. More serious is that direct and diffuse PAR fractions where not observed (a sensor for direct/diffuse shortwave radiation exist and is not very expensive). This study modelled direct/diffuse PAR from the calculated cloudiness index and observed total PAR using the empirical equations (eq 8-12) by Reindle et al 1990. These equations where parameterized in coastal northern (US and Europe) locations, which is a very different than the climate type, latitude and elevation of the current study site (this site is at least 1500 m above the highest site of Reindle). I therefore question the accuracy of the Reindle equations to this site.
Finally, all environmental variables covary, with strong diurnal and seasonal cycles. For example, if you repeat the analysis from table 1 but based on time of day (e.g., compare 7-9 am to 12-2 pm) in the summer you will find very different alpha, Pmax and Rd. The point here is that the affects you attribute to more diffuse radiation, could be actually the effects of lower temperature or higher humidity. Your analysis approach does very little to disentangle the covarying drivers of transpiration and photosynthesis. Your path analysis confirms it (without actually solving the problem).
Especially for a single-site study, which is not generalizable from the start, getting more depth in the analysis of the hypothetical effects and linking the observations better to current models for the effects of direct/diffuse radiation is critical, and missing from this study.
Other, easier to address comments:
How are you calculating Reco (eq. 2) ?
Table 1 and table 2 do not indicate any form of uncertainty (goodness of fit? RMSE?)
"Please contact the corresponding author to access data" is not a valid data availability statement. Please publish the half-hourly dataset of meteorological and flux observations. Preferably through ChinaFlux.