Palaeohydrological studies on closed-lakes (lakes without an outflowing river) make a lot of sense. I did my PhD on this type of system in Ethiopia. When the precipitation/evaporation ratio is high, lake level is high and salinity is low, and vice versa.
I’m much more dubious about palaeohydrological studies on open lakes — “wide places in a river”. Can lake-level changes be reconstructed in such such systems, and can any changes be attributed to changes in precipitation/evaporation ratios? Of course, given sufficient changes in precipitation/evaporation, lakes that are now open can lose their outflow, become closed and have large fluctuations in lake depth and salinity. This probably happened to the North American Great Lakes in the early Holocene.
Some of the proxies used to reconstruct lake level in open lakes are doubtful, for example the water lily remains used by Stolze et al. Assemblages of diatoms, chironomids and cladocera have been used to reconstruct lake depth using transfer functions. None of these reconstructions are likely to be very robust as depth per se is not an ecologically important variable. Instead, depth is correlated with several ecological factors such as temperature, oxygen and light availability. If these correlations are not stationary over time, then reconstructions may be spurious. This is the message of Steve Juggins’ “sick science” paper, and that of a paper I wrote with Gaute Velle and others testing chironomid-inferred lake-depths reconstructed using a within-lake transfer function.
I would be much more convinced by palaeohydrological studies if they used hydrological models to demonstrate that the lake-level changes they infer are possible given plausible changes in precipitation, temperature or other hydrologically important variables. Such models are commonly used for closed-lakes — I have one in my PhD thesis — but I haven’t found one in any of the palaeoecological papers on open lakes I have looked at.
Most of the hydrological modelling work I’ve found on lakes concerns large lakes, but I have found one paper making predictions of hydrological changes under different climate-change scenarios.
Yao, H., Scott, L., Guay, C., & Dillon, P. (2009) Hydrological impacts of climate change predicted for an inland lake in Ontario by using monthly water balance analysis. Hydrological Processes, 23: 2368-2382.
Yao et al study Harp Lake, a <1km2 lake in the boreal ecozone in Ontario, Canada, that is not too dissimilar from the small lakes analysed by many palaeoecologists. They examine two climate-change scenarios for 2050 – one warmer and wetter, taken as the continuation of current trends, and one warmer and drier, taken from a general circulation model. With the second scenario, the mean annual temperature is over 3°C warmer than modern, a change roughly comparable with the difference between the pre-industrial and the Early Holocene thermal maximum. The annual precipitation difference relatively small, 20 mm or 2%.
This large temperature change and consequently large change in evaporation is predicted to cause a mean lake-level change of 2 cm. Two centimetres.
The paper discusses why the lake-level response is so small — far smaller than any palaeoecological study could ever hope to reconstruct — concluding that the response time of the lake is very short. When it rains, the lake level rises quickly in response to increasing inflows, and this lake-level rise causes the outflow stream to increase discharge, which quickly returns the lake to its stable level after the rain ceases.
Harp Lake is not an analogue for all palaeohydrological studies on open lakes. Some studies are on much larger lakes that will have much longer residence times, and so larger potential for climate change to change lake level. Other studies are on lakes that have other lakes upstream, making their hydrological response more complex.
Yao et al’s finding that the level of an open lake can remain virtually unchanged despite large climate changes makes me sceptical of papers interpreting changes in lakes levels in small open lakes as being driven by precipitation/evaporation changes. Other factors need to be considered, including the robustness of the reconstruction, and other hydrological changes, for example in the amount of coarse woody debris in the outflow streams. For small lakes in Europe (except Ireland) and North America, I wonder what the effect of beaver dams on lake level might be?