Subpermafrost groundwater systems: Dealing with virtual reality while having virtually no data

SummaryStudies on the Earth’s hydrology in general thrive on abundant data, while data on certain groundwater systems are virtually absent as a result of their inaccessibility. This poses challenges for understanding and modeling such systems, yet modeling is often the only option to study them. When it comes to limited data availability, a simple model may have a better predictive performance than a complex model. In a case study of a subpermafrost groundwater system on Svalbard we applied three simplified models which do not cover all processes, and compared their outcomes. The results of all models were different and sometimes contrasting. By combining all model results within their associated assumptions we show how dynamic processes in data-limited subpermafrost groundwater systems can be interpreted. Using multiple interpretations of a system by making different assumptions can thus be used to understand processes in data limited groundwater systems. Our results also illuminate the fierce data scarcity of subpermafrost groundwater systems, and the necessity of resolving this.

► We discuss modeling approaches for data-limited systems.
► We compare three highly simplified models for a subpermafrost groundwater system.
► Despite the models’ contracting outcomes, we reject none.
► Instead we review these multiple outcomes within their associated simplifications.