Surface Energy Balance Closure at Different Temporal Scales in the WFIP2 HRRR Model

Measurements of the turbulent and radiative surface energy fluxes were made in the Columbia River Basin during the Second Wind Forecast Improvement Project (WFIP2). Previous work studied the partial surface energy budget (SEB) closure based on the half-hourly, daily, monthly, seasonal, and sub-annual (~10-month) averaged data collected over different underlying soil surfaces (dry, wet, and frozen). Here, we analyse the High-Resolution Rapid Refresh (HRRR) model to assess the full closure produced by its land surface model over the same time scales. Observed half-hourly direct measurements of the partial SEB show that the sum of the turbulent sensible and latent heat fluxes systematically underestimates positive net radiation by around 20-30% during daytime, and overestimates negative net radiation at night. Though the model shows similar imbalances, it contains the missing ground fluxes not included in the observations. Furthermore, partitioning of the radiation and turbulent flux components show biases, which offset in the budget. However, on average, the observed imbalance is significantly reduced at longer averaging timescales, and the partial SEB can be closed for this site within reasonable limits on seasonal and sub-annual timescales. The model SEB will be analysed for its imbalances over the same averaging time scales, and compared over wet, dry, and frozen soil surfaces, and the role of the ground flux will be determined. Individual SEB components will be analysed, to identify specific errors in the land surface model, and how these errors may be reduced by longer averaging intervals.