Well,

The basic laws of groundwater hydraulics are linear. Thus, in problems depending

on several factors, the influence of each factor may be calculated apart, and the results added. This principle, that of superposition, will be the guideline throughout the theory. It makes understanding easy and allows complicated calculations to be split up into elementary ones.

Regards

Exactly. For example, a watershed with a 5 km2 lake. When simulating with 2×2km resolution, the model will calculate specific hydrological processes in the lake. Conversely, when doing simulation with 5×5 km resolution, the lake will be homogenized as a adjacent land use, eg., forest. In this case, with/without lake will definitely affect simulated results.

In the Groundwater (GW) Remediation Industry we face this challenge regularly, particularly in the forms of heterogeneity and preferential flow. For me it is a critical factor to understand the type of data that you need to derive. For example, if you are developing a water supply system then flow and permeability over large volumes is best measured as an average and GW engineering principals rule. That is the high permeability or hydraulic conductivity zones are measured and averaged over long (thick sections of aquifer). In these cases production of GW is all about the bulk production capacity. In this case there is very little need to understand the relationship between low and high permeability zones that may be present. In these cases use of engineering based methods measurements are very accurate and fairly predictive. However, when cleaning up groundwater the scale is much smaller and the zones that matter are much thinner and often more transmissive than the surround "bulk" of the aquifer. Thus, measuring an average of the entire aquifer section generally leads to an entirely different (inaccurate)set of results than is actually present in the zones that matter. What is becoming clear over time is that groundwater features that remediation folks are interested in are set by sedimentologic processes at the time of deposition. These deposition processes that are scaleable (that same features are present in micro-scale and macro-scale) and are very predictable. From my work to date, when you can gain a reasonable level of understanding of the underlying geologic -sedimentologic processes that deposited the aquifer (rock or soil) then factor in the hydrogeologic"details" the scale effects you mention are much more understandable and actually become somewhat predicable.

In hard rock aquifers we often made the experience that a single investigation measure (pump test in a borehole, packer test in a borehole section...) demonstrates values much too high for a long term draw-down (e.g. by a tunnel), even if well storage and other effects are considered during interpretation of the results. By interpolation of the results on the entire planned underground construction (considering the geological model, obviously), the calculated discharge seems much too high and often exceeds the recharge by precipitation in the entire catchment. The reduction of the prediction, then, starts to be an assumption only and does not base anymore on the test results. We do have, therefore, an important scale problem between the borehole test and the later underground excavation. The reasons for that can be different and, as mentioned by earlier answers, linear effects in the aquifer. We made the experience that groundwater recharge and ductile behavior of rock mass have a stronger influence on the inflow than the fragmentation of the rock mass itself.

Hi Peter, I suggest reading one of my papers about scale effects in karst aquifers on this link: Article The Karst Permeability Scale Effect of Sete Lagoas, MG, Brazil

On the references, you can find other citations about this topic. Best!