Atmospheric rivers and extreme precipitation in the western U.S under climate change
Atmospheric rivers (ARs) are the key driver of extreme precipitation in the western, and they have profund influence on the water resource availability in this region. Under the projected climate change, ARs are projected to be more frequent and stronger. To investigate the impact of such change from ARs on local precipitation, we conducted a high-resolution WRF simulation over the whole western U.S. The model is configured at grid spacing of 6 km to better resolve convective systems and topographic effect. The historical climate between 1981-2015 was reconstructed from the NARR-driven simulation. Regarding the future climate, we selected 5 CMIP5 models, and took a pertubation approach to simulate the climate between 2041-2017 under psuedo warming. The analysis of the simulation is going on, and the first part that checks the quantitative relationship between ARs and extreme precipitation has come out (Chen et al. 2018). Please feel free to check it out!
Reference: Chen et al. 2018
Probable maximum precipitation in a changing climate
Current engineering practice ensures the safety of large or critical infrastructures (such as dams or nuclear waste disposal sites) by designing them against Probable Maximum Precipitation (PMP), a theoretical upper bound of extreme precipitation. Advances in the climate studies allow us to estimate future PMPs under the projected climate change. To communicate between future PMP changes and existing infrastructure safety, consistent historical PMP estimation is required. Here we propose a hybrid approach that combines both climate model data and engineering tradition to evaluate the future extreme precipitation risks (Chen et al., 2017b). In this study, five earth system models (ESMs) with good performance on the atmospheric rivers in the US Pacific Northwest (PNW) region were selected to estimate the PMP based on the 1970-2016 simulation (left panel). Comparison of the ESM-based PMP with conventional PMP (as estimated from HydroMeteorogical Reports, HMRs) indicates good agreement (right panel). Ensemble estimates provide better consistency, and they would reveal the uncertainty in the PMP estimation (such as the orange and blue bars in the right panel, which indicate the std and range, respectively). This allows us to check the statistical significance of the future change in PMP. Using this approach, we found that PMP by 2099 in the PNW region will significantly increase by 50% of current level. For the details of this approach, you are welcome to check out my paper.
Reference: Chen et al. (2017b).
The relationship between extreme precipitation and meteorological conditions
Traditionally, PMP is estimated through moisture maximization, which inexplicitly assumes that precipitation is mostly sensitive to atmospheric moisture availability. To reasonably maximize the observed precipitation, it is necessary to know how the meteorological factors are limiting the storm magnitude. Using ERA-Interim (shown in the figure below) and North America Regional Reanalysis (NARR) products, we checked the roles of vertical air motion, moisture availability (precipitable water), and atmospheric instability (CAPE). Each of them dominates the extreme storms in different regions. In general, at the middle-latitude regions, precipitation is more limited by the vertical wind speed. In practice, this suggests that vertical wind (or horizontal convergence) may be the major factor to consider when we modernize the PMP estimation across the US.
To find out the feasibility of numerical models, such as WRF, in PMP estimation, we also did extensive studies over historical extreme precipitation events. Most of the previous efforts focused on the reconstruction of storm events in the recent decades. However, in the engineering practice, much older storms (since the 1910s) also provide valuable information during the infrastructure risk evaluation. Therefore, it is necessary to find out how old storms we can comfortably reconstruct by now. In the study of Chen and Hossain (2016) and Chen et al. (2017a), we explored the optimal WRF configuration for atmospheric river-induced precipitation simulation. We found that model performance is most sensitive to boundary condition choice, then to the parameterization options. In spite of this, some parameterization schemes perform better across various boundary conditions and model grid size, suggesting they might be prioritized in the simulation of similar events.
Carbon fluxes in the northern high latitude wetlands
Carbon fluxes (CO2 and CH4) in the northern high latitude wetland ecosystem plays an important role in the climate system.
Reference: Chen et al. (2015)