Optimizing shrub parameters to estimate gross primary production of the sagebrush ecosystem using the Ecosystem Demography (EDv2.2) model

Abstract

Abstract. Gross primary production (GPP) is one of the most critical processes in the global carbon cycle, but is difficult to quantify in part because of its high spatiotemporal variability. Direct techniques to quantify GPP are lacking, thus, researchers rely on data inferred from eddy covariance (EC) towers and/or ecosystem dynamic models. The latter are useful to quantify GPP over time and space because of their efficiency over direct field measurements and applicability to broad spatial extents. However, such models have also been associated with internal uncertainties and complexities arising from distinct qualities of the ecosystem being analyzed. Widely distributed sagebrush-steppe ecosystems in western North America are threatened by anthropogenic disturbance, invasive species, climate change, and altered fire regimes. Although land managers have focused on different restoration techniques, the effects of these activities and their interactions with fire, climate change, and invasive species on ecosystem dynamics are poorly understood. In this study, we applied an ecosystem dynamic model, Ecosystem Demography (EDv2.2), to parameterize and predict GPP for sagebrush-steppe ecosystems in the Reynolds Creek Experimental Watershed (RCEW), located in the northern Great Basin. Our primary objective was to develop and parameterize a sagebrush (Artemisia spp.) shrubland Plant Functional Type (PFT) for use in the EDv2.2 model, which will support future studies to model estimates of GPP under different climate and management scenarios. To accomplish this, we employed a three-tiered approach. First, to parameterize the sagebrush PFT, we fitted allometric relationships for sagebrush using field-collected data, gathered information from existing sagebrush literature, and borrowed values from other PFTs in EDv2.2. Second, we identified the five most sensitive parameters out of thirteen that were found to be influential in GPP prediction based on previous studies. Third, we optimized the five parameters using an exhaustive search method to predict GPP, and performed validation using observations from two EC sites in the study area. Our modeled results were encouraging, with reasonable fidelity to observed values, although some negative biases (i.e., seasonal underestimates of GPP) were apparent. We expect that, with further refinement, the resulting sagebrush PFT will permit explicit scenario testing of potential anthropogenic modifications of GPP in sagebrush ecosystems, and will contribute to a better understanding of the role of sagebrush ecosystems in shaping global carbon cycles.

Characterizing Dryland Ecosystems Using Remote Sensing and Dynamic Global Vegetation Modeling

“…The second assumption is that remote sensing GPP products capture the dynamics of vegetation production in our study area. This work builds on lessons learned from previous efforts to estimate GPP in the same study area Pandit et al, 2019;Renwick et al, 2019), all of which recommend additional investigations into model development within drylands.…”

“…However, an understanding of DGVM capabilities and limitations in an ecosystem is necessary to capture uncertainties. An evaluation of a DGVM should include model parametrization, sensitivity analysis (SA), calibration, and evaluation (Fer et al, 2018;Keenan et al, 2013;Kuppel et al, 2012;Pandit et al, 2019;Post et al, 2017;Renwick et al, 2019;Santaren et al, 2007;Wang et al, 2001). There is an information gap regarding DGVM evaluation in drylands and more specifically in regions where the drivers in ecosystem processes may vary across an elevation gradient.…”

“…This approach may be limiting in ecosystems in which GPP capacity and activity are controlled by different processes across a study area . Many studies have investigated the correlation between simulated GPP and GPP estimated from EC towers or remote sensing (Antonarakis et al, 2014;Pandit et al, 2019;Renwick et al, 2019;Trugman et al, 2016), which is more applicable to GPP capacity rather than activity. More explicit metrics of GPP activity are required to fully capture the performance of models in terms of vegetation dynamics.…”

“…For the sensitivity analysis (SA) and model calibration, we utilized shrub PFT parameters with strong relationships to GPP simulation (Table 3.2). The initial and lower/upper values are based on literature review or derived from Pandit et al (2019). A complete description of these parameters is presented in Pandit et al (2019), and broader descriptions of PFT parameters for ED is available in Longo et al (2019b).…”

“…The initial and lower/upper values are based on literature review or derived from Pandit et al (2019). A complete description of these parameters is presented in Pandit et al (2019), and broader descriptions of PFT parameters for ED is available in Longo et al (2019b). In this study, we implement a global sensitivity analyses method known as the Morris method (Morris, 1991).…”