Current Paleoclimatic Research

Our agent-based model for settlement and culture change in Southwest Colorado from A.D. 600-1300 is based in part on a series of paleoproductivity data planes, one for each of these 700 years. These dataplanes in turn are calculated from a network of middle-elevation tree-ring chronologies that contain climatic information. As we have gained more experience with these dataplanes over the first months of project activity, we have come to realize that they contain little in the way of low-frequency change. That is, they capture very accurately changes in productivity from year-to-year, but because of the way the underlying tree-ring chronologies were constructed, long-term trends (many decades or centuries in length) are much more poorly expressed. Records of vegetation change reconstructed from pollen deposited in lakes or bogs are an excellent source for reconstructing the low-frequency component absent from these tree-ring records. Therefore, in summer 2003, we revisited and are reanalyzing a site in the La Plata Mountains of Southwest Colorado first studied in the 1970s and known to have a high-resolution record applicable to our area, which we will be able to link with the tree-ring chronologies to reinsert in them the low-frequency component they currently lack.

Multiple directly comparable distinct sources of proxy data for climate history and vegetation responses are seldom available, or even collected at comparable scales. Moreover, they do not necessarily respond to the same elements of climate change. Our team faces such a situation in modeling the relationships between 700 years of environment and culture changes on the Colorado Plateau. We already employ a very useful archive of tree-ring data and will collect more to tune the information directly applicable to constructing a more precise paleoproductivity model. Still, with all their advantages, tree-ring indices do not reveal individual plant association responses to immediate or long-term environmental trends. Tree-ring responses to climate changes may only suggest how plant communities, individual species or individuals may have responded, whereas the actual local demise of species, reshuffling of plant communities, or shifts in elevation of upper or lower tree lines must come from the plant remains themselves.

With appropriate methods, pollen and macrofossils from lake cores may mesh well with and add paleoenvironmental interpretations to tree-ring reconstructions of climate. The ability to do so, however, depends on the accuracy of a chronology developed for reconstructing pollen deposition rates. The innovative aspect to this proposed expansion of our research is that we intend to sample the cores continuously in uniform time slices rather than by arbitrary depth slices. Specifically, we know enough about the bog to be cored to know that we will be able to analyze slices of the cores that incorporate 10 years of deposition (or some other similarly short period, the exact length of which will be determined after the deposition rates are calculated from the AMS ¹⁴C ages). See Figure 1 for an example of this technique applied in Oregon. Pollen influx rates (pollen/cm²/yr) in these samples and identification of macrofossils will reveal the regional and local pollen rain. Ratios of pollen from this site (for example, the spruce/pine ratio; see Figure 2) have been shown to be sensitive to changes in precipitation, including long-term changes in precipitation. If low-frequency changes in precipitation emerge from this analysis, as we suspect they will based on previous work (e.g., Petersen 1988), we will use these ratios, computed every 10 years, to renorm the tree-ring chronologies to which they can be linked.

We already know that the top 1.5 m of sediment at this site (Beef Pasture), accumulated during the last 2000 years. By recollecting these sediments, and by obtaining ample AMS ¹⁴C dates, and by employing current close-interval palynological techniques (Mehringer and Wigand 1990; Figure 1), we will realize a more direct comparison of fossil pollen (buttressed by plant macrofossils) with regional tree-ring indices filtered to correspond to the time-lengths of the pollen samples. Two fossil pollen study sites (Twin Lakes and Beef Pasture) located in landslide-dammed depressions at different elevations on the same mountain gave Petersen (1988) an opportunity to approximate the local vegetation history.

Figure 1. Distribution of 67 pollen samples
spanning the last 500 radiocarbon years,
Diamond Pond, Harney County, Oregon.
Pollen samples can be calibrated to tree-ring
dates with high temporal precision.
From Mehringer and Wigand 1990: Figure 13.13.

Because of differing Engelmann spruce growth response to climate variation at its lower and upper elevational limits, the proximity of these two radiocarbon-dated pollen records not only allowed their correlation, but also suggested paleoclimatic interpretations doubtfully derived from either alone.

In developing a Paleoenvironmental sequence for Southwest Colorado, Petersen first analyzed a 3.75 m, 9800- year core from Twin Lakes (3290 m elevation) near the upper limit of Engelmann spruce. Twin Lakes is currently located 250 m below the average elevation of timberline in open spruce-fir forest. Although this site is reasonably well dated (nine radiocarbon dates), deposition has been relatively slow. The last 2000 years -- the period of most interest to Puebloan archaeologists -- falls within the upper 45 cm. It does, however, allow estimates of summer temperatures derived from fluctuations of upper tree-line. Fortunately, Beef Pasture (3060 m elevation), 230 m below and 5 km west of Twin Lakes, holds within a 3.90 m core 5500 years of accumulated sediment, roughly twice the deposition rate of Twin Lakes. This more rapid deposition enhances the prospect for obtaining a high-resolution reconstruction of vegetation changes in the last 2000 years.

Beef Pasture is an open grass-and-sedge meadow surrounded by mixed aspen, Douglasfir, and spruce and fir forest (Figure 2).

Figure 2. Coring at Beef Pasture, La Plata mountains, summer 2003.
From l to r: Sander Kohler (Pullman High School), Tim Kohler,
Dave Johnson (Washington State University), Ken Petersen (University of Utah).

Numerous overlapping cores, and 12 radiocarbon dates established the chronology. However, only one of these dates falls within the Pueblo period of interest (Figure 3). This is the reason that we revisited and recored this site, emphasizing only those portions of its record that date to our study (A.D. 600-1300). The relatively few samples from this period in the previous study are nevertheless sufficient to suggest the probability that there is a long-term (low frequency) downward trend in precipitation from ca. A.D. 600-920, a sharp rebound until the early-mid 1100s, and a sharp decline after that. These apparently strong long-term trends appear only faintly in the (high-frequency) tree-ring records on which our paleoproductivity model is based.

Figure 3. Spruce/pine ratio at Beef Pasture, La Plata Mountains, for A.D. 500-1300

(Petersen 1988). Values below 60 indicate climatic conditions drier than those of the last

110 years in the La Plata Mountains and surrounding region. This sequence is built on 23

samples; we will rebuild it based on >100 samples, calibrated to 10 tree-ring-corrected

radiocarbon ages.

Village Home