Australia: The Land Where Time Began

A biography of the Australian continent 

Lake Mixing regimes – Worldwide Alteration in Response to Climate Change

Much of the liquid freshwater of the Earth is held in lakes, which support biodiversity and are the providers of key ecosystem services to people around the world. They are vulnerable to climate change, however, e.g., through shorter durations of ice cover, or through the rise in temperature of the surface water of lakes. In this study a 1-dimensional lake model is used in order to assess the impacts on mixing regimes in 635 lakes around the world of climate change. The lake model was run in this study using data from 4 state-of-the-art model projections of 21^st century climate under 2 emission scenarios. Many lakes are projected to have reduced ice cover under the scenario with the higher emissions (Representative Pathway 6); it is projected that about ¼ of lakes that are ice covered seasonally will be permanently ice free by 2080-2100. It is projected that surface waters will warm, with a median warming across lakes of about 2.5^oC. About 100 of the lakes that were studied are projected by the simulations of this study to undergo changes in their mixing regimes. Of these 100 lakes about ¼ are classified currently as monomictic – undergoing 1 mixing event in most years – and will become systems that are permanently stratified. About ⅙ of these are currently dimictic, i.e., mixing twice per year, and will become monomictic. Woolway & Merchant concluded that many lakes will mix less frequently in response to climate change.

Among the changes in lakes related to climate is ice cover for shorter durations in winter (Sharma et al., 2019; Magnuson et al., 2000; Fang & Stefan, 2009; Magee et al., 2016) and higher temperature of lake surface water (Schneider & Hook, 2010; Magee & Wu, 2017; O’Reilly et al., 2015; Austin & Colman, 2007; Livingstone, 2003). It has been shown by recent global studies of the temperature of lake surface water, predominantly lakes that experience seasonal ice cover, are warming at rates that exceed ambient air temperature (O’Reilly et al., 2015; Austin & Colman, 2007). Understanding of the consequences of warming on lake ecosystems has been improved by studies of temperature response of lakes to climate change (O’Reilly et al., 2003; O’Beirne et al., 2017). In this paper Woolway & Merchant assessed how projected climate trends are likely to change the stratification and mixing of 635 lakes that are distributed globally. It is critical that stratification and mixing are considered in order to understand the repercussions of the change in temperature throughout lake environments and associated ecosystems, as these aspects of lake dynamics exert substantial control of the stratification of nutrient fluxes, oxygenation and biogeochemical cycling (North et al., 2013; Yankova, Neuenschwander, Köster & Posch).

Stratification and mixing regimes in lakes

Thermal stratification of lakes occurs as a result of the thermal expansion properties of water. The balance between turbulence, which acts to enhance mixing, and forces of buoyancy, which act to suppress turbulence and result in vertical layering (Boehrer & Schultze, 2008), determine the time evolution of stratification. Strong control of the transport of nutrients and oxygen between the surface and deep water of lakes and the vertical distribution and composition of Lake Biota, is exerted by the vertical layering that exists during stratification. Lakes that are permanently mixed (continuous cold/warm polymictic) or those that mix frequently (discontinuous cold/warm polymictic) differ markedly in their physical, chemical and biological function from lakes that are permanently stratified (oligomictic, characterised by variable temporal periods of mixing that is incomplete, interspersed with occasional mixing) (Boehrer & Schultze, 2008; Boehrer, von Rohden & Schultze, 2017). Seasonally stratifying lakes can be classed as dimictic if they have 2 stratification seasons, or monomictic (cold or warm) if they stratify only once per year. Seasonal mixing serves as a basis for classifying lake regimes (Lewis, 1983) and a necessary component of projecting how lake ecosystems will respond to climate change.

The mixing class of a lake depends primarily on whether or not it experiences ice cover annually and the number of times in a year which it stratifies continuously. Ice-covered lakes have a tendency to occur in high latitudes and regions of high elevation that are less maritime regions. The climatological duration of ice cover varies systematically with mean air temperature, which is illustrated by satellite observations of 635 lakes, of which ~50% experienced ice cover annually. With regards to the stratification criterion for mixing class, observations of surface water temperature can be used to distinguish between lakes that are monomictic and dimictic (warm or cold) lakes: in warm monomictic lakes, the temperature of surface water does no drop below 4^oC (near the maximum density of freshwater), whereas the temperature of the surface water does not warm above 4^oC in cold monomictic lakes. There is no threshold water surface temperature that separates thermally that are stratifying and polymictic lakes, as there are other factors that can have a substantial influence, particularly depth of a lake (shallow lakes can mix easily). It is suggested by the global heterogeneity of lake sizes and depths (Dee et al., 2011; Verpoorter et al., 2014; Messager et al., 2016) that lake mixing classes should be distributed heterogeneously.

Global patterns of lake mixing regimes

Woolway & Merchant assess in this study the mixing class of 635 lakes around the world by the development of a scheme of classification applied to numerical simulation results from a lake model, and the use of this model to project future mixing classes under climate change scenarios. This approach enables meromictic, oligomictic, dimictic and polymictic mixing classes to be determined.

In order to assess the contemporary mixing classes key parameters of the lake model, Flake (Mironov, 2008; Mironov et al., 2010), are first optimised to represent the dynamics of each individual lake. The model is constrained by the optimisation to represent lake surface water temperatures time series that are observed. Comparison of simulations with independent temperature observations, ice cover and regimes of lake mixing under historic climatic conditions are used to evaluate the ability of the optimised lake model to represent a wide range of lake dynamics. Good agreement was obtained. The lake model is capable of accurately simulating historic multidecadal variations in lake temperature back to the beginning of the 20^th century as well as identifying successfully the mixing regimes of 72% of lakes for which independent mixing regime classifications were found, which supported its utility for multidecadal projections. The lake model that was forced with representations of historic climatic conditions explained up to 85% of Interdecadal lake surface temperature variability.

The regimes of lake mixing that were identified demonstrate a diverse array of mixing types. In the global dataset used in this study dimictic lakes are most common, which is the result of the majority of lakes in this study being large (all lakes studied had an area of greater than 27 km², and were situated to the north of 40^oN, where the global abundance of lakes is at its highest (Verpoorter et al., 2014). In the northern temperate latitudes the proportion of dimictic and polymictic lakes is large, as was expected, and in the tropics meromictic lakes are common.

Climate-related changes in lake mixing regimes

The lake model is forced by 4 climate projections available from the Inter-sector Impact Model Intercomparison Project (Frieler et al., 2017), namely, HadGEM2-ES, GDFL-ESM2M, IPSL-CM5A-LR and MROC5, under 2 Representative Pathway (RCP) scenarios, in order to project changes in the mixing class in the future. In this paper Woolway & Merchant show or quote the spread of results from the lake model across all 4 climate model projections, to indicate the uncertainty of projections. Changes that have been projected for 2080-2100 are quoted relative to the period 1985-2005.

The responses of lake mixing regions to climate change are complex and may not be closely associated with change in any one climatic variable. Instead, the mixing regime of a lake will depend on a change of climatic factors that contribute to the lake heat budget, such as air temperature, solar and thermal radiation, cloud cover wind speed, and humidity. The results of this study project that under future scenarios RCP 2.6 and 6.0, the number of annual ice covered days will substantially decrease by 2080-2100. The average decrease for RCP 2.6 is 15 days (across all lakes that are covered by ice seasonally during the historic period) and the standard deviation of this mean change across the 4-member ensemble is 5 days. The projected mean change for RCP 6 is -29 ± 8 days. Under RCP 6 the most extreme cases, the decrease that has been projected in ice-covered days is more than 60 days. It is projected by the simulations that by the end of the 21^st century under, RCP 6.0, 24±5% of lakes that were covered by winter ice in the historic period will be free of ice. It is projected that the increase in mean lake surface temperature will be 1.1 ± 0.4 and 2.3 ± 0.6^oC under RCP 2.0 and 6.0, respectively, by 2080-2100. Projected warming for individual lakes can be higher, the largest increase that has been projected is 5.4 ± 1.1^oC under RCP 6.0; 99 ± 0.5% of lakes are projected to experience higher mean temperatures under RCP 2.6, increase under RCP 6.0.

Decreases in winter ice cover and increases of lake surface temperatures would be expected qualitatively to modify the distribution of lake mixing regimes. Next, they investigated the global extent and magnitude of response in the projections to quantify this expectation. It was suggested that alterations in the lake regimes will occur during the 21^st century. Specifically, in the projections under RCP 2.6 and 6.0, 59 ± 7 and 96 ± 15 lakes change mixing class, respectively.

A change from warm monomictic to meromictic is the most common alteration in mixing class identified (25 ± 5% of altered lakes under RCP 6.0). According to this a substantial minority of lakes that do not experience and stratify once annually are projected to become stratified systems permanently by the end of the 21^st century. Additionally, all of the lakes that have been identified as being oligomictic during the historic period transition of meromictic class by 2080-2100. By the end of the 21^st century a lack of vertical mixing will result in a reduction in upwelling of nutrients from deep to shallow waters and a reduction of oxygen concentrations in deep water which can lead to reduced lake productivity (O’Reilly et al., 2003) and the formation of dead zones (North et al., 2014) in deep water, respectively. Oxygen depletion at depth can be detrimental to the habitat for fish (Regier, Holmes & Pauly, 1990) as can modify biogeochemical processes that results in, e.g., the potential release of phosphorus and ammonium to the water column (Mortimer, 1941) and the release of metal ions that are potentially toxic (Davion, 1981).

A change from dimictic to warm monomictic (17 ± 5 of lakes that have been altered under RCP 6.0) is the 2^nd most common alteration in mixing class that has been identified (across the model ensemble). According to Woolway & Merchant this alteration occurs when lakes that were historically ice covered no longer freeze in winter, though they continue to stratify in summer. There are implications of the projected absence of winter ice for those lake ecosystems, which includes, among other factors, changes in water quality (Weyhenmeyer, Westöö & Willén, 2008) and the production and biodiversity of phytoplankton (Weyhenmeyer, Bleckner & Patterson, 1999).

There are a very small number of lakes that have been altered that are projected to experience fewer continuous periods of stratification and, therefore, to transition from dimictic to polymictic mixing classes. Changes in any of the meteorological drivers that act at the surface of the lake can result from such an increase in mixing. An important role in lake stratification and mixing can be played, e.g., by short-term variations in the speed of surface wind and could nudge some lakes into a different mixing regime (Woolway et al., 2017). The influence of changes in wind speed would, however, be expected to have the most pronounced effects on the mixing regime of shallow lakes. All of the lakes that have projected to undergo an increase in the number of mixing events per year are among the shallowest 1% of the lakes that were included in this study. There are other factors that were not considered in this study that may also be important for the alterations in mixing regimes in specific lakes, such as input of groundwater (Rosenberry et al., 2015), increase of input of cold water from retreating glaciers (Peter & Sommaruga. 2017), thermal pollution from nuclear plants (Kirillin, Shatwell & Kasprzak, 2013) and changes in the magnitude of influent water (Valerio et al., 2015), which will be of particular importance for lakes with short residence times or extensive variations of lake level (Rimmer et al., 2011). The mixing regime of lakes can also be influenced by changes in lake transparency (Shatwell, Adrian & Kirillin, 2016), though are not expected to be a dominant driver in alteration of mixing regime in large lakes, such as those that were included in this study. It has been shown that the influence of transparency on mixing and stratification with increasing size of lakes (Fee et al., 1996; Read & Rose, 2013) and the vertical thermal structure would be constrained strongly by fetch (Gorham & Boyce, 1989). The transparency is expected, therefore, to have a greater influence on of the vertical thermal structure of lakes that are relatively small compared with lakes with larger ones, as was investigated in this study.

According to Woolway & Merchant scattered evidence has been found of alterations in mixing regimes already taking place, and the ecological consequences of these changes are beginning to appear (Kainz, Ptacnik, Rasconi & Hager, 2017; Ficker, Luger & Gassner, 2017). The projections of future lake mixing regime alterations that have resulted from this study span a wide range of locations, sizes and climatic contexts, and suggest a complex pattern of lake responses to climate change. Climatic conditions interact with lake-specific contexts, geomorphology in particular, which results in geographical distribution in lake mixing alterations being heterogeneous. The projections do not support simple expectations of regional consistency in the responses of lakes whereby lakes in a given region will change in a similar manner. Changes in lake temperature will not always translate to changes in lake mixing regimes: some of the lakes that are projected to experience the highest degree of surface warming are not projected to undergo change in their mixing class. Most lakes that have been projected to alter their mixing regimes have been found to currently display anomalous behaviour relative to their dominant mixing classification in some years. Specifically, in ⅔ of lakes that had been projected to experience alteration in their mixing regimes there were at least 3 years when they had anomalous mixing regimes during the period 1985-2005. Seasonally ice-covered lakes that were classified as dimictic at present, though they also experienced some winters when they were free of ice, have been projected to become predominantly monomictic by the end of the 21^st century. The projections, therefore, that the intuition that these lakes that presently exhibit anomalous years relative to their mixing class are more likely to transform to a different mixing regime in the future.

Sources & Further reading

Woolway, R. I. and C. J. Merchant (2019). "Worldwide alteration of lake mixing regimes in response to climate change." Nature Geoscience 12(4): 271-276.