For over half a century, researchers at Hubbard Brook Experimental Forest (HBEF) have been assembling one of the most comprehensive long-term ecological data sets in the world. Early work at HBEF provided seminal insights into forest nutrient cycling and water budgets, eventually informing the creation of the Brook90 model—a lumped conceptual framework still in use today. Yet, as hydrologists have begun to seek more detailed, site-specific representations of water movement, the community has looked toward fully coupled, physics-based models.
A new study—Soil Texture-Based Parameterisation and Hydrological Insights of a Fully Coupled Surface and Subsurface Model at the Hubbard Brook Experimental Forest, USA—adds to this line of research by applying Latin hypercube sampling (LHS) to calibrate a computationally heavy, physics-based model for eight headwater catchments in HBEF. According to the paper, “Parameterisation of fully coupled integrated hydrological models is challenging,” and brute-force calibration often proves infeasible because of high simulation times. By leveraging LHS, the authors systematically explored parameter combinations (soil characteristics, evapotranspiration factors, and snowmelt rates among others) in far fewer total runs than a traditional trial-and-error approach would have required.
The resulting model achieved an average Nash–Sutcliffe Efficiency (NSE) of 0.80 for simulating hourly discharge—an important benchmark that indicates strong agreement with observed streamflow data across multiple sub-watersheds. Notably, the authors saw a 7% improvement in NSE after adding snowmelt and evapotranspiration parameters, underscoring the importance of winter–spring transitions for runoff generation in northern hardwood forests. While earlier conceptual frameworks at Hubbard Brook, such as Brook90, provided valuable “big-picture” estimates of streamflow and evapotranspiration, the new study extends that work to capture the heterogeneity of soil layers and the influence of bedrock, steep slopes, and seasonally variable conditions.
Much of the value here comes from tying these improved simulations directly to the rich, long-term data sets for which Hubbard Brook is famous. Over the past decades, researchers have meticulously recorded precipitation, temperature, and streamflow alongside measures of forest growth, soil chemistry, and nutrient fluxes. By calibrating the fully coupled model within this historically well-studied setting, the team has woven new, high-resolution hydrological insights into the fabric of Hubbard Brook’s ecological narrative. The authors highlight that the model’s refined parameters—especially those dictating infiltration and subsurface storage—could eventually help interpret larger ecological phenomena, including shifts in nutrient export and forest resilience under changing climate scenarios.
In sum, this work illustrates a thoughtful convergence of long-term ecological monitoring with emerging modeling techniques. By applying LHS to narrow down parameter sets for a detailed, physics-based model, the study achieves both computational efficiency and a robust fit to decades of observational data. It represents a continued step in translating HBEF’s long-term records into predictive capacity—valuable both for the scientific community and for forest management under increasingly variable environmental conditions.
Norouzi-Moghanjoghi, K., Fakhraei, H., Valipour, M., & Driscoll, C. T. (2025). Soil texture-based parameterisation and hydrological insights of a fully coupled surface and subsurface model at the Hubbard Brook Experimental Forest, USA. Hydrological Processes, 39(1), e70045. https://doi.org/10.1002/hyp.70045