An analysis of multi-decadal profiles from 11 stratified Adirondack lakes revealed a generally consistent pattern of response that included warming and thinning of the epilimnion, cooling and expansion of the hypolimnion, and declining hypolimnetic dissolved oxygen. Among stratified lakes, the increase in epilimnetic temperatures during 1994–2021 (0.27 ± 0.07 oC per decade) was smaller than the increase in regional annual air temperature during this period (0.35 ± 0.16 oC per decade). The average trend for Adirondack lakes is somewhat smaller than global lake surface water trends reported by O’Reilly et al. (2015; 0.34 oC per decade for 1985–2009) and Pilla et al. (2020; 0.36 oC per decade for 1970–2010). Jane et al. (2021) report an accelerated trend in lake temperature during 1990–2000, followed by a “warming hiatus” during 1998–2009 (Winslow et al. 2018). The 12-y warming hiatus spans 43% of the 28-y time series for Adirondack lakes and may account for the overall lower warming rate reported here. Winslow et al. (2017) showed that warming rates among Wisconsin lakes also varied seasonally, and that these tracked seasonal patterns in air temperature trends. Adirondack air temperature trends indicate that greatest warming occurred in shoulder months (May and October), but our lake data were largely restricted to June-August profiles, precluding estimation of trends outside this period. Lake trends based on June-August data may underestimate overall warming during the stratified period.
Adirondack lake trends were consistent with prior work documenting epilimnetic warming and hypolimnetic cooling (Bartosiewicz et al. 2019). Though some studies have reported hypolimnetic warming (e.g., Dokulil et al. 2006), significant hypolimnetic warming was observed in only one of the stratified lakes in this study. Hypolimnetic cooling has been attributed to loss of water clarity, which has a shielding effect on deeper layers as more heat is retained in the surface layer. Water clarity is decreasing among Adirondack lakes, particularly those that are recovering from acidification (Bukaveckas 2021). As chemical recovery has progressed following reductions in acidic deposition, optical recovery has manifested through increasing lake color (brownification). The combined effects of atmospheric warming and decreasing light penetration has strengthened thermal stratification and altered the depth distribution of heat via epilimnetic thinning and hypolimnetic expansion.
Hypolimnetic oxygen depletion is common even among oligotrophic lakes (Nurnberg 1995; Mathews and Effler 2006; Bidanda et al. 2018; Yuan and Jones 2020), with higher values typically associated with increasing trophic state (e.g., as indicated by chlorophyll-a and TP; Rippey & McSorley 2009). Foley et al. (2012) report volumetric hypolimnetic oxygen depletion values for Blenhalm Tarn of 131 to 252 µg L− 1 d− 1, which they consider to be representative of eutrophic conditions. Biddanda et al. (2018) reported that volumetric hypolimnetic oxygen demand in Muskegon Lake was high and comparable to other temperate eutrophic lakes based on a range of values 71–148 µg L− 1 d− 1. Rippey and McSorley (2009) report values among eutrophic lakes reaching 240 µg L− 1 d− 1. Adirondack lakes fall in the lower range of values for hypolimnetic oxygen depletion (median = 46.8 ± 6.1 µg L− 1 d− 1; range = 22.1 to 82.1 µg L− 1 d− 1) as would be expected given their low productivity. However, values in the range ascribed to eutrophic lakes were not uncommon (observed in 8 of 11 lakes). The occurrence of high rates of hypolimnetic oxygen depletion in these oligotrophic lakes complicates the use of oxygen depletion metrics as indicators of excess nutrient enrichment (e.g., in regional and national lake assessments). For the Adirondack dataset, attempts to explain inter-annual variation in oxygen depletion based on climate variables (e.g., monthly or seasonal temperature anomalies) and in-lake variables (chlorophyll-a as an indicator of lake productivity) were not successful, leaving an open question as to factors driving inter-annual variation in hypolimnetic oxygen deficits. There is some evidence that inter-annual variation in the timing of the onset of stratification may account for variation in oxygen deficits, as earlier stratification results in greater cumulative oxygen depletion (Ladwig et al. 2021; Jane et al. 2022). Woolway et al. (2021) suggest that decreasing water clarity favors faster spring warming of the surface layer and earlier stratification, which may account for the observed declines in hypolimnetic oxygen resources of Adirondack lakes.
Concurrent changes in the depth distribution of heat and light availability may also influence deep water oxygen resources via effects on phytoplankton photosynthesis. Deep chlorophyll maxima are common in Adirondack lakes and elsewhere where photic depths exceed mixing depths (Noges et al. 2011; Leach et al. 2017). Among the lakes included in this study, photic depths were typically 5–10 m, and often exceeded mixing depths. Deep chlorophyll maxima in these lakes are associated with oxygen supersaturation and carbon dioxide minima (P. Bukaveckas, unpubl. data), potentially indicating elevated productivity in deeper layers (Wilkinson et al. 2015). The influence of phytoplankton photosynthesis on hypolimnetic oxygen conditions was apparent in cases when successive monthly profiles showed increases in hypolimnetic oxygen (i.e., negative oxygen depletion rates). These were commonly observed (~ 25% of profiles) in some lakes (Indian, Limekiln, Piseco and Raquette). Decreasing water clarity would negatively affect photosynthesis in deeper layers, and may potentially contribute to observed declines in hypolimnetic oxygen. The effects of decreasing water clarity may be partially offset by the concurrent thinning of the epilimnion, whereby chlorophyll maxima persist, but occupy a shallower position in the water column. We have previously documented that light attenuation in Adirondack lakes recovering from acidification has increased from 0.42 ± 0.08 m− 1 to 0.64 ± 0.04 m− 1 (average = 0.23 ± 0.07 m− 1) during 1988–2018 (Bukaveckas 2021). This corresponds to a decrease in the depth of the photic zone (z 1%) of 130 cm per decade, which exceeds the rate of epilimnetic thinning among lakes in this study (70 ± 8 cm per decade). These findings suggest that photic depth is decreasing faster than mixing depth resulting in reduced light availability below the mixed layer. Diminishing photosynthesis in deeper layers may be a contributing factor to the observed declines in hypolimnetic oxygen.
Whereas factors driving inter-annual variation in hypolimnetic oxygen depletion were difficult to discern, inter-lake variation was largely explained by differences in bathymetry. Prior studies have used various metrics in relating hypolimnetic oxygen conditions to lake basin shape (e.g., Nurnberg 1995). Here we used the ratio of hypolimnetic sediment area to hypolimnetic volume as an indicator of layer thickness. Highest rates of oxygen depletion were observed among lakes with thin hypolimnia, highlighting the importance of sediment vs. water column oxygen demand in determining hypolimnetic oxygen resources. Bouffard et al. (2013) reported similar findings for eutrophic Lake Erie where hypolimnetic oxygen depletion decreased from ~ 700 µg L− 1 d− 1 in shallow stratified areas (hypolimnion < 2 m) to < 200 µg L− 1 d− 1 in deeper areas (hypolimnion thickness = 3–8 m). Our findings suggest that lake bathymetry is a key factor determining their sensitivity to climate effects on deep water oxygen resources. Lakes with thin hypolimnia exhibit higher rates of oxygen depletion, and are more prone to anthropogenic effects that further diminish dissolved oxygen. Our conception of how individual lakes differ in their response to a regional stressor (climate warming) recalls earlier work documenting their response to changes in the chemistry of atmospheric deposition. In both cases, individual lakes showed varying degrees of response to a regional-scale stressor (acid deposition, climate warming) due to site-specific factors. For the former, it was shown that depth of till in the watershed was a key determinant of sensitivity to acidification, as the acidity in runoff was largely neutralized in lakes with thicker till deposits (Goldstein et al. 1985). For the latter, bathymetry determines sensitivity to climate effects on hypolimnetic oxygen because lakes with thinner hypolimnia are prone to low oxygen conditions that are exacerbated by earlier onset and stronger stratification. Identifying lake- and watershed- specific factors determining sensitivity to stressors is valuable to understanding variable responses among lakes within a region.
Changes in the depth distribution of light, heat and dissolved oxygen have important implications for lake biogeochemistry and food webs. Lake carbon balances are affected by changes in the temperature and thickness of thermal layers because lake temperature is linked to microbial activity and C mineralization (Bartosiewicz et al. 2019). Trends in Adirondack lakes suggest that potentially greater C mineralization due to warming of epilimnia and their underlying sediments may be offset by greater C storage in deeper sediments as cooling and expansion of the hypolimnion results in a smaller proportion of the lake bottom being exposed to warm temperatures. Fish habitat is affected by changes in depth gradients of temperature and dissolved oxygen, particularly for thermally-sensitive species such as salmonids (Arend et al. 2011; Kraus et al. 2015). Brook trout are an iconic species of the Adirondacks and throughout their native range in the eastern United States and Canada. The combined effects of epilimnetic warming and hypolimnetic oxygen depletion compress their habitat to a narrower range of depth layers with suitably low temperature and sufficient dissolved oxygen. Effects on habitat are responsive to both climate warming and recovery from acid deposition. The changes in habitat observed among our study lakes are consistent with those documented during lake liming experiments. A focal point for these studies was Woods Lake, where a series of lake and watershed limestone additions transformed the lake from an acidified (pH < 5), clearwater, marginally stratified and well-oxygenated state, to a circumneutral condition of reduced water clarity, strong thermal stratification and low hypolimnetic dissolved oxygen (Bukaveckas and Driscoll 1991a,b). Based on current trends, we predict that deep water oxygen resources of Adirondack lakes will continue to diminish, in some cases falling below the desired threshold for cold water fisheries. In this way, climate change will reduce habitat for brook trout, even as lakes and populations recover from acidification (Robinson et al. 2010; Warren et al. 2016).
Lastly, we have focused largely on the effects of climate change and variability on stratified lakes, though we recognize that shallow, unstratified and marginally stratified lakes are common in this and other lake regions (Eilers and Selle 1991; Downing et al. 2006). Marginally stratified and transitional lakes showed an increase in strength of thermal stratification and decreasing dissolved oxygen, but their response was more variable. Shallow lakes have received less attention in studies of climate effects, though recent work suggests that warming is associated with the occurrence of transient summer stratification, which can result in rapid oxygen depletion of bottom waters and concomitant effects on fish distribution (Sondergaard et al. 2023).