Abstract:

Terrestrial ecosystems are important systems for mitigating greenhouse effects and regulating climate change. One of the main purposes of carbon cycle research in terrestrial ecosystems is to determine the value of the net CO₂ exchange (NEE) between atmospheric and terrestrial ecosystems. To estimate the amount of net ecosystem CO₂ exchange, the Vegetation Photosynthesis-Respiration Model (VPRM) with calibration for different ecosystems was introduced. This study was based on flux observation data, remote sensing data and meteorological data obtained from the observation tower of the Tuotuo River during the growing season (from May to September 2019) to calibrate the four VPRM parameters, including maximal light use efficiency (λ), half saturation value of photosynthetically active radiation (P₀), and two vegetation parameters (α, β). In addition, the simulation effect of the model was verified using three schemes. The results showed that 1) considering the absence of photosynthesis in plants at night, the night NEE data were entirely characterized by ecosystem respiration. Using the night NEE and air temperature data of the Tuotuo River during the growing season in 2019, the respiratory parameters α and β were calculated by linear fitting, with values of 0.034 μmol·m⁻²·s⁻¹·℃⁻¹ and 0.217 μmol·m⁻²·s⁻¹, respectively. Subsequently, the ecosystem respiration R during the day was calculated by α, β, and daytime temperature, which was used to estimate the gross ecosystem CO₂ exchange (GEE) using subtracting ecosystem respiration (RE) from NEE during the day. The values of λ (0.119 μmol) and P₀ (64.920 μmol·m⁻²·s⁻¹) were calculated using the mathematical equation between GEE and gross primary productivity (GPP) as well as their corresponding parameters T, W, P, E, and P₁. 2) The accuracy of the model was further checked at a scale of 30 min and day-night. The value of the regression equation slope (R²=0.462) was calculated to be 0.566 between the simulated NEE (using the VPRM model) and the observed NEE from the flux station at a time scale of 30 min, with an average root error of 0.805 and a mean bias of 0.561. The simulated NEE showed instability on a time scale of day or night, which was higher or lower than the observed NEE. However, NEE was simulated well by the VPRM at the day-night scale (R²=0.713), with a regression equation slope of 0.918, an average root error of 0.473, and a mean bias of 0.342. 3) On the day-night scale, the interpolation of missing values was further analyzed based on the VPRM model and REddyProc packages (R language) using the collected data from the flux stations. The missing values in the data input between the VPRM and REddyProc programs exhibited a high correlation (day: R²=0.934; night: R²=0.975) and the interpolation precisions of the two methods were similar. Compared to the day scale, the VPRM model is more suitable for the night scale. 4) Two important parameters in the VPRM model were photosynthetically active radiation and temperature, and the values of these two parameters differed greatly on rainy, cloudy, or sunny days. Therefore, the influence of different weather conditions on the NEE simulation was analyzed with the result that the VPRM model had the highest applicability in sunny conditions (R²=0.829, RMSE=0.346, MAE=0.267). In summary, the VPRM model with calibrated parameters is highly available in the typical alpine desert steppe ecosystem, which not only provides a foundation for regional NEE estimation but also provides a new idea for estimating the missing value of ground observation data.

Key words: alpine desert steppe, vegetation photosynthesis and respiration model (VPRM), net ecosystem CO₂ exchange (NEE), flux, photosynthetically active radiation (PAR), parameter calibration, model verification

摘要：

为实现利用植被光合呼吸模型（VPRM模型）在陆地生态系统估算大气与陆地生态系统之间净CO₂交换量，需要针对不同生态系统进行VPRM模型参数的率定。基于位于典型高寒荒漠草原的沱沱河站2019年生长季（5—9月）的通量观测数据，结合遥感数据和气象数据，对VPRM的4个关键参数[最大光能利用率λ、光照为半饱和条件下光合有效辐射值P₀和植被参数（α、β）]进行了率定，并对模型模拟效果进行了验证。结果表明，1）典型高寒荒漠草原生态系统VPRM模型的关键参数α、β、λ和P₀值分别为0.034 μmol·m⁻²·s⁻¹、0.217 μmol·m⁻²·s⁻¹、0.119 μmol、64.920 μmol·m⁻²·s⁻¹。2）VPRM模型在昼夜时间尺度上模拟效果最好，模拟值与观测值散点图的斜率接近0.918（R²=0.713），均方根误差为0.473，平均误差为0.342。3）VPRM模型和R程序包（REddyProc包）插补缺失值后两组数据相关性较高（昼R²=0.934；夜R²=0.975），两种方法插补精度类似。4）参数率定后的VPRM模型在晴天适用性最好（R²=0.829），均方根误差和平均误差均较小，分别是0.346和0.267。该文通过本地化参数率定和模拟效果验证，得到了典型高寒荒漠草原生态系统的VPRM模型，且具有较好的估算效果；该结果不仅为开展区域NEE估算奠定理论基础，而且还为地面观测数据的缺失值估算提供了新的思路。

关键词: 高寒荒漠草原, VPRM模型, NEE, 通量, 光合有效辐射, 参数率定, 模型验证