A new publication will be coming out in the Canadian Journal of Fisheries and Aquatic Sciences. This work, funded by a grant I received from NASA / AK EPSCoR, combined LiDAR derived maps of biomass, disturbance exposure modeling (landslides and wind) and various topography metrics to predict stream chemistry in over 40 headwater streams in the Heen Latinee Experimental Forest.
The results suggest that steam chemistry, at least for some species, is fairly predictable in these systems. In general, slope was the primary driver of carbon and nitrogen concentrations, which corresponds to a several things- productivity, species composition, and Sphagnum/muskeg concentration being the most significant. For some chemical species, disturbance exposure seemed to be important, particularly phosphorus. This makes sense if you think about disturbance as exposing rocks for weathering.
We need to expand before we can go broad scale, getting more alpine-only drainages and drainages dominated by alder (though alder was a component). But it's a good step towards modeling non-glacial stream chemistry based on the characteristics of the watersheds that they inhabit.