Yanjun Gan

Dr. Yanjun Gan is a Research Scientist II at NOAA’s Physical Sciences Laboratory (PSL) and is affiliated with the Cooperative Institute for Research in Environmental Sciences (CIRES) at the University of Colorado Boulder (CU Boulder). Prior to joining PSL/CU Boulder, Dr. Gan worked at the University of Texas at Arlington from 2019 to 2024, first as a Research Engineering Scientist and later as an Assistant Professor of Research. From 2015 to 2019, he was with the Chinese Academy of Meteorological Sciences (CAMS), where he began as an Assistant Research Scientist and was later promoted to Associate Research Scientist.

Dr. Gan's research focuses on hydrometeorology, water resources, land data assimilation, uncertainty quantification, and related fields.

1. Model Parameter Uncertainty Quantification
   The accuracy of model predictions is significantly influenced by uncertainties stemming from data errors and model deficiencies. Quantifying and reducing uncertainties in model parameters can greatly enhance the reliability of these predictions. Dr. Gan developed a systematic framework for uncertainty quantification, combining stepwise sensitivity analysis with adaptive surrogate-based multi-objective optimization (Gan et al., 2014, 2015, 2018). This approach improves the practical assessment and reduction of model parameter uncertainties.
2. Model Parameterization Uncertainty Quantification
   The ability to select from different parameterization schemes within the same modeling framework allows for a range of physical configurations, as seen in models like WRF and Noah-MP. However, this flexibility complicates the selection of an optimal configuration. To address this, Dr. Gan employed the uniform design method to minimize the number of experiments needed to represent different parameterization combinations. He also utilized ANOVA and multiple comparison tests to identify dominant processes and optimal configurations for various hydroclimate, soil, and vegetation conditions (Zhang et al., 2016; Gan et al., 2019, 2024; Li et al., 2024).
3. Forecast-Informed Reservoir Operation
   Reservoirs play a crucial role in mitigating flood and drought risks in many urban areas across the US. Managing the temporary use of flood storage effectively requires reliable forecasting of reservoir inflow and the use of appropriate guide curves. Dr. Gan developed a flood forecasting and warning system to improve streamflow prediction (Gan et al., 2017; Zeng et al., 2020). He also optimized seasonal guide curves for a system of reservoirs in Texas, aiming to enhance water supply reliability while minimizing the risk of flooding (Gan et al., 2024).
4. Snow Data Assimilation
   Accurate and timely estimation of snowpack is vital for water resource management and flood preparedness in regions with significant snow cover. Errors in the model representation of snowpack have been a major source of uncertainty in streamflow predictions in these areas. To address this, Dr. Gan developed an algorithm that blends satellite and in situ snow water equivalent (SWE) data, producing a reliable gridded SWE product for the CONUS (Gan et al., 2021). He then assimilated this product into the National Water Model (NWM) to improve snow and streamflow predictions in mountainous regions (Gan et al., 2022).