Regional and local patterns of soil nutrients at Rocky Mountain treelines

Abstract

The soils across treeline should vary because of direct effects of biological differences of coniferous subalpine forest and the herbaceous alpine tundra in Colorado. In addition, the change in life form may indirectly affect soils because of interactions of the vegetation and wind-driven deposition processes. This is particularly important as nitrogen (N) saturation is a growing concern in high elevation ecosystems, and treeline is predicted to be a deposition hotspot. The vegetation transition at treeline provides an opportunity to test the effects of vegetation, topography, and external inputs on soils at three spatial scales. First, a regional evaluation of soils at eleven abrupt treeline sites was made comparing sites on east and west aspects both east and west of the Continental Divide (CD). Second, soils were compared in the adjacent forest and tundra. Finally, edge effects were assessed along transects spanning treeline. At the regional scale, total soil N was higher east of the CD and on east aspects while exchangeable calcium was higher on east aspects and at sites west of the CD. Higher lead (Pb) concentration in the forest organic horizon was associated with lower ²⁰⁶Pb/²⁰⁷Pb ratios, an indication of greater anthropogenic Pb inputs. However, the spatial pattern in soil Pb suggests a different source area or transport mechanism than N. Within individual sites, the soils differed between the forest and tundra in almost every measured variable, but edge effects were minimal on both sides of these abrupt treelines. While a direct link between the observed soil patterns to deposition of external inputs cannot be made based on the study design, the observed soil patterns suggest that the impacts of acid deposition are amplified or attenuated by processes such as dust deposition.

Introduction

The effects of plants on soil heterogeneity have been well documented at small spatial and temporal scales. Plants alter the horizontal and vertical distribution of nutrients in soils (Zinke, 1962, Steltzer and Bowman, 1998, Lovett et al., 2004). The mere presence of plants in a matrix of bare ground creates heterogeneity in soils (Schlesinger et al, 1996). At longer time scales, organisms, especially plants, are one of the state factors that shape pedogenesis (Jenny, 1941). While there are relationships between biomes and soil orders at the global scale, other state factors almost inevitably vary at these large spatial scales. Of all the state factors, the community of organisms is typically the most difficult to control; thus, biosequences (sites where vegetation is the only state factor that changes) are difficult to find (Jahren, 2004). One prominent biosequence that has been identified in various areas of the northern hemisphere is the “prairie-timber transition zone” (Jenny, 1941). While the proximal causes of this abrupt vegetation shift have been debated for almost a century (e.g. Gleason, 1922), the differences in soil organic matter concentration, pH, C:N ratio, and clay distribution have been well documented (Jenny, 1941). The soils at alpine treeline, a similar grassland–forest transition, have received comparatively little attention.

The unique setting of alpine treeline, in that it tends to occur in windy areas with high topographic relief, suggests the tree–wind interactions could create predictable patterns in inputs and soil properties throughout the transition between forest and tundra. Unequal snow distribution based on tree-wind interactions below treeline is the most visually striking example of this effect (Hiemstra et al., 2002). A less obvious effect is the difference in soil properties throughout the vegetation transition associated with the presence of trees. Studies near treeline have largely focused the effects of the highest individual trees on soils in a matrix of alpine tundra (Cairns, 1999, Seastedt and Adams, 2001, Shiels and Sanford, 2001). Dramatic differences in soil horizonation and chemical properties have been observed under individual trees (Holtmeier and Broll, 1992). Individual trees have been dated up to 1000 years old and the location of treeline is suggested to have been stable for thousands of years (Ives and Hansen-Bristow, 1983). On Niwot Ridge, Colorado the differences in soils upwind and downwind of individual trees near treeline were as great as the differences between the alpine tundra and subalpine forest soils (Liptzin and Seastedt, 2009). Gradual changes in soil properties related to gradual changes in state factors are well documented in mountainous regions such as the Front Range of Colorado (e.g. Birkeland et al., 2003). However, soils at treeline may vary at much shorter spatial and temporal scales because of both the indirect effects of trees interacting with the wind and the direct effects of the change in vegetation at the abrupt forest edge.

Abrupt treeline transitions are also conducive to evaluating the effects of forest edges on soils. In general, forest edges, especially in coniferous forests, have been found to be efficient at capturing atmospheric deposition and pollution (Weathers et al., 2001). The Landscape Continuum Concept (LCC) posits that treeline should be subject to particularly high external inputs as trees should intercept materials transported across and scoured from the extensive areas of treeless alpine tundra by the high winds consistently from the same direction (Seastedt et al., 2004). While the LCC did not make an explicit prediction about the spatial relationship between forest and tundra, the amount of transported material should be greatest when the forest is situated downwind of the tundra. Further, the LCC does not explicitly predict the lateral extent of the accumulation zone below treeline. While a theoretical framework for quantitatively assessing ecological boundaries, such as the forest edge at treeline, is developing rapidly, it remains deceptively difficult to quantify the location and size of transition zones (Cadenasso et al., 2003). A variety of techniques have been used to delineate edges, such as non-linear regression models (Cadenasso et al., 1997, Toms and Lesperance, 2003) and the depth of edge influence (Chen et al., 1995). In most cases the vegetation edges studied are due to forest fragmentation and not naturally occurring which makes can conflate the vegetation effects with the disturbance effects.

At larger spatial scales, the interaction between the forest edge at treeline and regional differences in the sources of materials transported in the atmosphere may also create variability soils. For example the spatial patterns of total nitrogen (N), exchangeable calcium (Ca), and total Pb, which have been linked to acidic, dust, and total deposition respectively, may vary regionally in treeline soils because they differ in their sources and how they are entrained and transported in the atmosphere. Nitrogen pollution originating from fossil fuel combustion and agriculture east of the Front Range is a growing concern in high elevation areas of Colorado (Williams et al., 1996, Baron, 2006). There is growing evidence that N has altered foliar chemistry and soil nutrient cycling in the subalpine forest east compared to west of the Continental Divide (Rueth and Baron, 2002). Measurements of atmospheric deposition, soil properties, and lake sediments have all suggested that eolian deposition carried from proximal and distal sources to the west provides a crucial source of base cations such as Ca, to nutrient poor high elevation soils in Colorado (Thorn and Darmody, 1980, Caine, 1986, Litaor, 1987, Muhs and Benedict, 2006). While small dust inputs occur every year to the region (e.g. Liptzin and Seastedt, 2009), there are also large regional events where dust originating in the desert southwest blankets the mountains of Colorado (Rhoades et al., 2010). Lead and Pb isotope ratios have been used, mostly in the northeastern U.S., to assess overall patterns of atmospheric deposition (Weathers et al., 2000, Kaste et al., 2003). Changes in sediment Pb concentrations related to human activities have been documented in high elevation lakes in Colorado, but the source area and spatial pattern of Pb deposition on the landscape is not well established (Baron et al., 1986).

We used the existing patterns in treeline soils at three spatial scales to test predictions of the direct and indirect effects of the vegetation on soils. Specifically, our goals were to 1) Evaluate the regional patterns of N, Ca and Pb, in addition to other commonly measured soil properties, in treeline soils. Nitrogen was predicted to be highest east of the Continental Divide because of the proximity to the source area whereas all three elements are predicted to be greater on east aspects based on the predictions of the Landscape Continuum Concept; 2) Use the vegetation transition at treeline as a biosequence to examine the effects of vegetation type on soils; 3) Estimate the lateral extent of the forest edge and the magnitude of edge effects at this naturally occurring forest edge.