Results at Last!
Here is one of the long overdue RESULTS reports that Brian mentioned recently. This one addresses our efforts to use subfossil remains of midge larvae to investigate Late Glacial temperatures at Big Pea Porridge Pond. As you will see, it is a work-in-progress. But this report will bring you up to date at least.
If you are interest in investigating a particular phenomenon in nature, you search the landscape for a setting that makes viewing that feature clear. To study distribution patterns of animals, for example, you don’t start out with the bewildering complexity of a tropical rainforest. On the other hand, if you are primarily interested in unraveling the story of a particular location, you must apply available procedures and make the best of the results you get. I’ll use our studies using chironomid (non-biting midge) larvae as a proxy (an indicator) of past climate temperatures to illustrate this point. Today, these insect larvae are abundant and diverse in lake bottom sediments. The same was true in the past.
To study past temperature changes, paleoecologists have selected small, shallow water bodies because they are sensitive to overlying atmospheric conditions. These water bodies mix easily and frequently with any modest wind. While in any water body, direct heating by the sun may be limited to surface waters, wind-driven mixing in shallow pools circulates warmed water throughout the water body and in the process heats up bottom sediments throughout the pond as well. When climate-sensitive adult midges reproduce and lay their eggs in such ponds, the eggs sink to the bottom where they hatch into tiny larvae that feed on organic debris, unicellular plants, and in some cases on each other. As benthic (i.e., bottom dwelling) creatures, they pass through 4 successive molts before eventually becoming pupae and finally adults. Figure 1 shows examples of these larvae.
Long after the soft parts of such larvae rot and disappear, their hard head capsules remain intact. For every egg laid or adult hatched, there are 4 hardened, exoskeleton head capsules (inset in Figure 1) left behind in the bottom sediments. Each year, debris from the plankton community, leaf litter, pollen grains, etc. sinks to form the next layer of bottom sediments, trapping beneath them sediment layers from previous years. Each year these settled sediments include midge head capsules representing the variety of these creatures alive at that time.
As waters circulate within a pond, in a fashion similar to that seen when you gently swirl a bowl of water containing small particles, there is a tendency for small items such as the heads of midge larvae to be swept from the shallows around the perimeter towards its center, where they eventually sink through the water and join that year’s layer of bottom sediments. Since there is no corresponding mechanism to sweep such midge remains from the deeper bottom toward the shallow shoreward regions, bottom samples taken toward the pond’s deep center are most apt to include the full array of chironomid types available at that time – both the shallow (often warmer-water tolerating ones) and the deeper (more cold tolerant types). Also in the deep center, the delicate annual layering of bottom sediments is far less apt to become disrupted by water movements or by biological activity (burrowing mussels, wading moose, or human boating or swimming traffic). It is conventional wisdom that coring “the deep spot” provides the least disturbed, most representative chironomid samples.
From our sediment core, I took 3 cc subsamples at 10 cm intervals starting at the bottom of the core and progressing towards the surface. Refer back to Scientific Posting 22 for a description of subsampling techniques. For each subsample examined, we determined the “species composition” of the community of chironomids alive at that location and at that period in time.
Many of the chironomids found in deeper sections of the sediment core (i.e., from several 1000 years ago when local conditions were much colder) represent species no longer active in today’s warmer climate at PPP. However, these very same species are alive today in more poleward, colder localities such as ponds of northern Canada, Labrador, etc. Scientists have visited these locations and studied the environmental conditions in which these representatives of these species thrive today. Doing so has permitted them to determine that the mid-summer surface water temperature is the environmental variable that most accurately determines which species are present in bottom sediments and which are not. They have determined the optimal mid-summer surface water temperature required by each species. Using this information one can use a mathematical model (a partial least squares weighted average (PLS-WA) model) to blend the optimal temperatures for each species found in a given subsample, weighted by proportional abundance of that species in the subsample’s community, to produce the “best guess” inferred mid-summer surface water temperature that would fit a chironomid community composition matching that observed. For each subsample, a temperature point so generated is placed on a graph. Repeating this at 10 cm intervals up the sediment core generates a temperature profile for the period.
Studies of this sort in water bodies throughout northeastern North America have generated profiles similar to that shown in Figure 2. Starting to the right, at the retreat of glacial ice, temperatures rise steeply, but then shelve off, usually with a modest mid-shelf dip. This period is referred to as the “Killarney Oscillation” and in most cases occurs just older (i.e., deeper in the core) than 11,000 radiocarbon years before the present. Following the Killarney period, temperatures plummet to near glacial conditions (9-10 C) and they stay low for approximately 1000 years – from ca 11,000 – 10,000 radiocarbon years before the present). This period is known as the Younger Dryas period and may correspond to the time when glacial retreat was enough to permit meltwaters to, for the first time, flow out through the St Lawrence system, flooding the North Atlantic. This surge of frigid freshwaters may have been sufficient to block the Gulf Stream from delivering its included heat to northern North America and Europe. The Younger Dryas ends in a steep temperature build-back to something resembling today’s climate.
The top of Figure 3 shows an example of such a profile – this one generated from sediments taken at Splan Pond in New Brunswick. The Killarney Oscillation and Younger Dryas periods are obvious. The temperature profile generated from our analysis of the core taken at Big Pea Porridge Pond in New Hampshire may be seen in the bottom of Figure 3.
Comparing these two graphs, some overall similarities are seen, such as something generally similar to the Killarney Oscillation followed by a period of much lower temperatures. But there are also clear differences. 1) There is a spike in temperature late in what “should be” the cold Younger Dryas (marked as “4” on the Pea Porridge Pond figure). 2) There is a more modest overall temperature range observed (10-16 C at Big Pea Porridge Pond vs 0-18 C at Splan Pond). So how do we explain these differences?
Table 1 compares features at Big Pea Porridge Pond to those at 14 other locations where chironomid-based temperature estimates have been made.
Big Pea Porridge Pond lies furtherest to the south of all these sites, but not so much so as to account for the differences we found. Big Pea Porridge Pond is both the deepest of any of the sites (only Splan Pond at 10.4 m comes close) and by far the greatest in surface area of all (48 ha vs. a mean value of just 4 ha for the others). Although, Big Pea Porridge Pond is small compared to other New Hampshire lakes, it is large enough so that nearly all the input source of radiant heat from the sun is absorbed in its surface waters only. Looking progressively deeper In summer, this creates a layer of warm surface water and then a “thermocline” region of sharply falling temperatures overlying otherwise much colder deep waters. Such lakes are referred to as “thermally stratified”, and as such, resist vertical mixing. The creators of the PLS-WA model were clear that their model is designed for use on small, shallow water bodies that are not large enough to thermally stratify. Their model is designed for use in small, shallow, frequently mixed water bodies. But we were trying to learn about past conditions in our particular lake which doesn’t match this description. We are in fact applying a mathematical model to a circumstance beyond the model’s limits. But nonetheless, what can we learn about the past at Big Pea Porridge Pond?
Working in a larger, deeper, stratified pond means that only the shallow perimeter and the bottom sediments underlying its edges are significantly warmed by the sun, leaving the bulk of the lake’s interior and its underlying sediments much cooler and less variable in temperature. The published information on chironomids reveals that some larval types are associated with deeper, colder sediments (i.e., “profundal” sediments), while other species are limited to shallow, warmer sediments (i.e., “littoral” sediments). In Figure 4, we have calculated the percentage of each of the communities we observed at 10 cm intervals comprised of colder, profundal types. Sixty to eightly percent of our communities met this criterion. In previously published studies using chironomid larvae in this way, more than half the chironomids are warm littoral types. By studying communities from a deeper colder lake, we have observed communities more dominated by colder larval forms than is the case in ponds upon which the model is based. Therefore, our data matches low temperature patterns observed elsewhere, but it underestimates the higher temperature portion of the temperature profile, i.e., our smaller temperature range.
But the most obvious departure of the PPP temperature profile and that of Splan Pond is the late-Younger Dryas temperature spike. No one anywhere in the northeast has reported a similar late-YD thermal event at that time period. It makes you stop and think about the temperature conditions larvae living in the depths of thermally stratified PPP may have encountered. Eggs laid in PPP are the product of adults flying around in and totally dependent on the ambient climate conditions, including temperature. But in fact, once eggs have been laid in PPP and have sunk through the warm surface water layer followed by the thermocline, they enter cold deep water that, in a thermally stratified pond like this, shows very little temperature change year-round. Survival of larvae in all but the very shallow bottom sediments in this larger, deeper lake has little to do with temperature. Temperature is simply not much of a variable in their larval lives. What other conditions might influence which larvae survive and which do not in such lakes?
Frequently, in thermally stratified, non-circulating lakes, access to oxygen at depth is an issue. Oxygen is replenished in aquatic systems largely through diffusion into surface waters from the overlying atmosphere or as a by-product of light-driven plant photosynthesis – also a surface phenomenon. In small, shallow ponds with good mixing, frequent wind-driven circulation continually replenishes bottom sediments with oxygen. In deeper, non-circulating lakes, access to oxygen replenishment in deep sediments is a serious problem.
Once again, the published chironomid literature proved useful in providing information on which larval types are limited to conditions high in oxygen and which types are adapted to cope with reduced oxygen levels in bottom sediments. An example of an adaptation found in some species in the latter group would be the presence in body fluids of some species of hemoglobin which significantly improves oxygen uptake when it is scarce. Another low-oxygen adaptation is seen in species that build cylindrical burrows in bottom mud and use undulations to circulate water past their body. Comparison between species demanding high oxygen conditions vs those able to tolerate low oxygen levels vs those not distinguished by either characteristic, applied to communities observed from PPP, results in Figure 5.
One particularly striking change in proportions of these types can be found surrounding a depth of 790 cm within the core – exactly where the anomalous “temperature” spike occurred during the mid-Younger Dryas.
Our working hypothesis is that as the YD progressed, bottom oxygen in PPP declined. This resulted in a decline in high oxygen demanding species and a rise in those tolerant of lower oxygen. This represented an alteration in the community composition – in this case due to an oxygen crisis. But data reflecting this oxygen-related change was inserted into a mathematical model that only “understands” temperature. Consequently, the temperature model interpreted such a change in species present as necessarily being a result of a sudden temperature spike. Here is clear example of the problem in applying a model designed for use in environmental circumstances in which temperature change is the only or at least the primary controlling variable (e.g., small, shallow water bodies) to non-target circumstances in which two or more variables (in this case, both temperature & dissolved oxygen) are important.
It is possible that as the cold Younger Dryas period wore on, surface water temperatures may have dropped below 4 C. This is important because water is densest at 4 C. If less dense cold water (i.e., 1, 2, or 3 C) forms above deeper, denser, 4 C water, it will prevent the deep water from mixing with surface water. As a result of stagnation, dissolved oxygen used in bottom sediments would not be replenished. Oxygen levels would fall. Although year-round ice cover would also lead to stagnation, it would also block the input of chironomid eggs into the lake, and we noted no decline in chironomid larvae during this period.
We are currently exploring techniques for trying to distinguish between the roles of these two variables at PPP. If we are able to quantify changes in PPP’s chironomid community due to oxygen variations, perhaps we can “subtract” the oxygen influence to produce a clearer picture of temperature changes during this Late Glacial period.
Lee Pollock
