Abstract. Data-model integration plays a critical role in assessing and improving our capacity to predict ecosystem dynamics. Similarly, the ability to attach quantitative statements of uncertainty around model forecasts is crucial for model assessment and interpretation and for setting field research priorities. Bayesian methods provide a rigorous data assimilation framework for these applications, especially for problems with multiple data constraints. However, the Markov chain Monte Carlo (MCMC) techniques underlying most Bayesian calibration can be prohibitive for computationally demanding models and large datasets. We employ an alternative method, Bayesian model emulation of sufficient statistics, that can approximate the full joint posterior density, is more amenable to parallelization, and provides an estimate of parameter sensitivity. Analysis involved informative priors constructed from a meta-analysis of the primary literature and specification of both model and data uncertainties, and it introduced novel approaches to autocorrelation corrections on multiple data streams and emulating the sufficient statistics surface. We report the integration of this method within an ecological workflow management software, Predictive Ecosystem Analyzer (PEcAn), and its application and validation with two process-based terrestrial ecosystem models: SIPNET and ED2. In a test against a synthetic dataset, the emulator was able to retrieve the true parameter values. A comparison of the emulator approach to standard brute-force MCMC involving multiple data constraints showed that the emulator method was able to constrain the faster and simpler SIPNET model's parameters with comparable performance to the brute-force approach but reduced computation time by more than 2 orders of magnitude. The emulator was then applied to calibration of the ED2 model, whose complexity precludes standard (brute-force) Bayesian data assimilation techniques. Both models are constrained after assimilation of the observational data with the emulator method, reducing the uncertainty around their predictions. Performance metrics showed increased agreement between model predictions and data. Our study furthers efforts toward reducing model uncertainties, showing that the emulator method makes it possible to efficiently calibrate complex models.


Linking big models to big data: efficient ecosystem model calibration through Bayesian model emulation

Abstract. Surface fluxes from land surface models (LSMs) have traditionally been evaluated against monthly, seasonal or annual mean states. The limited ability of LSMs to reproduce observed evaporative fluxes under water-stressed conditions has been previously noted, but very few studies have systematically evaluated these models during rainfall deficits. We evaluated latent heat fluxes simulated by the Community Atmosphere Biosphere Land Exchange (CABLE) LSM across 20 flux tower sites at sub-annual to inter-annual timescales, in particular focusing on model performance during seasonal-scale rainfall deficits. The importance of key model processes in capturing the latent heat flux was explored by employing alternative representations of hydrology, leaf area index, soil properties and stomatal conductance. We found that the representation of hydrological processes was critical for capturing observed declines in latent heat during rainfall deficits. By contrast, the effects of soil properties, LAI and stomatal conductance were highly site-specific. Whilst the standard model performs reasonably well at annual scales as measured by common metrics, it grossly underestimates latent heat during rainfall deficits. A new version of CABLE, with a more physically consistent representation of hydrology, captures the variation in the latent heat flux during seasonal-scale rainfall deficits better than earlier versions, but remaining biases point to future research needs. Our results highlight the importance of evaluating LSMs under water-stressed conditions and across multiple plant functional types and climate regimes.

Modelling evapotranspiration during precipitation deficits: identifying critical processes in a land surface model

Inter-comparisons of mean, trend and interannual variability of global terrestrial gross primary production retrieved from remote sensing approach.

Calibration and validation of phenological models for Biome-BGCMuSo in the grasslands of Tibetan Plateau using remote sensing data

https://www.fs.fed.us/rm/pubs_journals/2021/rmrs_2021_dashti_h001.pdf

Performance of the Ecosystem Demography Model (EDv2.2) in Simulating Gross Primary Production Capacity and Activity in a Dryland Study Area

Process refinement contributed more than parameter optimization to improve the CoLM's performance in simulating the carbon and water fluxes in a grassland

The hydrodynamics and environmental factors in the Xiangxi River (XXR) and Shennong River (SNR), which are tributaries of the Three Gorges Reservoir (TGR), were monitored from July to August (the low water level period) and in October (the impoundment period) in 2018. The vertical distribution characteristics of chlorophyll a and other indicators of the two tributaries were analyzed during the different operation periods, and the factors that affected the vertical distribution in each period were discussed. The results showed that the vertical distribution of dissolved oxygen, water temperature, pH value, and chlorophyll a of the XXR and SNR during the low water level period was relatively consistent. The indexes 0-10 m (0-5 m for chlorophyll a) from the surface of the XXR and SNR, respectively, showed significant stratification and decreased with increasing water depth; the stability index of thermal stratification (RWCS/H) was 13.71-29.07 m-1, which was stable. After the water depth reached 10 m (5 m for chlorophyll a), the indexes tended to be stable along the water depth. During the impoundment period, there was no obvious stratification for each index; the stability index of thermal stratification was 0-0.5 m-1, the stability of the water body was weak, and the vertical variation of each index was relatively stable. The comprehensive trophic state index (TLI) of the XXR and SNR were 55 and 53 during the low water level period, respectively, indicating that they were in a slightly eutrophic state, and 39 and 46 during the impoundment period, respectively, indicating a mesotrophic state. Linear regression analysis showed that chlorophyll a, dissolved oxygen, water temperature, and pH in the two tributaries were significantly correlated in the vertical direction in the low water level period, indicating that dissolved oxygen, water temperature stratification, and pH were important factors affecting the vertical distribution of chlorophyll a. During the impoundment period, a large amount of backflow from the Yangtze River, a large fluctuation in tributary water level, and the decrease in RWCS/H were the important factors that affected the small vertical change in the water body. The enhancement of vertical mixing and the decrease in Zeu/Zmix were the key factors affecting the nutritional status of the water.

[Effects of the Three Gorges Reservoir Operation on Vertical Distribution of Chlorophyll a and Environmental Factors in Tributaries].

The arid region of northwest China is an extremely fragile area of natural ecology. With population growth and continuous expansion of urban scale, the ecosystem in the arid region is facing greater pressure. Scientific assessment and prediction of the value of ecosystem services in arid areas are necessary and of great significance for the sustainable development of regional ecological environments. In this paper, a parametric optimal geographic model is used to analyze the driving factors of ESV spatial dispersion in Jiuquan City as an example. The PLUS model is also used to simulate the spatial and temporal evolution of ESV in 2035 under the scenarios of natural development, urban development, water constraints and ecological conservation, based on the historical change pattern of land use and ecosystem service value (ESV). The results showed that from 1980 to 2020, the plowland, water and construction land area in Jiuquan City showed a significant increasing trend, while other land use types showed a decreasing trend. ESV has increased from CNY 139.394 billion to CNY 142.642 billion. The expansion of plowland and water area was the main reason for the increase in ESV. Elevation, temperature, and precipitation are the main driving factors of spatial differentiation of ESV in Jiuquan City. The interaction of natural and human factors enhances the explanatory power of each factor to the spatial differentiation of ESV. In 2035, the ESV in four development scenarios in Jiuquan City showed an upward trend, and the ESV of the ecological protection scenario was the highest, which was the best mode to realize a sustainable development in Jiuquan City in the future. This study can provide scientific basis and decision-making basis for Jiuquan City to formulate sustainable development strategy.

/pdf/multi-scenario-simulation-and-driving-force-analysis-of-1puju1su.pdf

Yali Li

Papers

Process refinement contributed more than parameter optimization to improve the CoLM's performance in simulating the carbon and water fluxes in a grassland

[Effects of the Three Gorges Reservoir Operation on Vertical Distribution of Chlorophyll a and Environmental Factors in Tributaries].

Multi-Scenario Simulation and Driving Force Analysis of Ecosystem Service Value in Arid Areas Based on PLUS Model: A Case Study of Jiuquan City, China