D.J. Pennock

Development and application of landform segmentation procedures

Landscape-scale approaches to research in soil science are explicitly focused on transfers of components within and between landscapes. Despite wide-spread recognition of the importance of these transfers, the application of landscape-scale approaches has been hindered by the lack of clear, reproducible research designs. Landform segmentation is used to divide natural and human-influenced landscapes into functionally distinct units. A specific type of landform segmentation, landform element classification, was used in a comparative mensurative design to compare the effects of cultivation on soil distribution and soil organic carbon (SOC) storage and in a manipulative design to determine the relationship between N2O emissions and fertilizer rate in a hummocky till geomorphic surface in southern Saskatchewan. Significant transfers of SOC and surface soil from convex shoulder units to lower slope positions occurred over the past 90 years, resulting in a change in the type of soils that occupy these positions at two research sites. The observed pattern is consistent with a tillage translocation dominated surface. The dominant control on N2O emissions in the landscape are spatial differences in water-filled pore space (WFPS) that are strongly controlled by water redistribution. Emissions from drier, shoulder landform element complexes are consistently low throughout the year, whereas a strong positive relationship between N fertilizer rate and N2O emissions occur in the wettest, level depressional elements.

Clear-cut forest harvest impacts on soil quality indicators in the mixedwood forest of Saskatchewan, Canada

Abstract The concept of soil quality is relevant to a range of human-induced disturbances on soil; our objective was to examine the impact of clear-cutting on soil quality conditions in six Mixedwood ( Populus tremuloides-Picea glauca ) stands in Central Saskatchewan, Canada. Soil quality conditions at two short-term (1 to 5 years) and four medium-term (6 to 20 years) clear-cut sites were compared to eight mature Mixedwood sites. All sites had a similar, albeit complex, distribution of soil and parent sediments. No major differences were observed between the short-term clear-cut sites and the mature Mixedwood sites. At the medium-term sites, substantial losses of soil organic carbon (24%), soil nitrogen (27%), and LFH thickness (28%) were observed, along with decreases in the soil surface (0 to 15 cm layer) of exchangeable calcium and magnesium (30%), soluble organic P (15%), and cation exchange capacity and base saturation (20%). The losses were much lower in the 15 to 45 cm layer of the soil. Overall, the levels of these soil components observed at the clear-cut sites were within the natural or undisturbed range as assessed at the mature Mixedwood sites; however the amounts of soil organic carbon and nitrogen at the clear-cut sites were lower than the natural range. Although the absolute levels of loss of soil organic carbon and nitrogen were much lower than in typical agricultural landscapes, the long-term ecological significance of the losses in these forested landscapes remains a concern.

Landscape-scale variability of N mineralization in forest soils

Abstract Our understanding of the controls on N-cycling and availability in forest soils following disturbance is limited. A comparative study was conducted to examine the spatial distribution of soil N in forest soils and assess the effects of site disturbance. Sampling grids were established within a 120 × 120 m representative area at a native site, and at recently (i.e. 4-year-old) burned and clear-cut sites. A three-dimensional classification of landscape form was used to stratify each landscape into distinct landform elements. The spatial distribution of inorganic-N was not related to landform element, irrespective of site disturbance, indicating an absence of topographic control at the scale studied. However, a narrowing of the NH4+-to-NO3− ratio at the clear-cut site compared to the native site suggests that N-cycling was influenced by site disturbance. Similarly, an increase in the size of the microbial biomass at the clear-cut site, coupled with a widening of the microbial biomass C-to-N ratio, suggest that disturbance altered both the size and composition of the microbial biomass. Potential N and C mineralization, and net nitrification in the forest floor and surface mineral horizons representing two distinct landform complexes were studied in a controlled aerobic 8-wk incubation experiment. Accumulation of NH4+ and NO3− differed markedly between sites although the effects of topographic position were generally non-significant. Inorganic-N accumulated principally as NH4+ in soils from the native site due to an extended lag in nitrification. In contrast, NH4+ accumulation in soils from the recently disturbed sites remained limited, whereas NO3− accumulation predominated. Thus, although topography did not markedly influence N distribution at the scale studied, site disturbance had a direct effect on N-cycling processes in these forest soils.

Testing the DNDC model using N2O emissions at two experimental sites in Canada

Measured data from two experimental sites in Canada were used to test the ability of the DeNitrification and DeComposition model (DNDC) to predict N2O emissions from agricultural soils. The two sites, one from eastern Canada, and one from western Canada, provided a variety of crops, management practices, soils, and climates for testing the model. At the site in eastern Canada, the magnitude of total seasonal N2O flux from the seven treatments was accurately predicted with a slight average over-prediction (ARE) of 3% and a coefficient of variation of 41%. Nitrous oxide emissions based on International Panel for Climate Change (IPCC) methodology had a relative error of 62% for the seven treatments. The DNDC estimates of total yearly emissions of N2O from the field site in western Canada showed an underestimation of 8% for the footslope landscape position and an overestimation of 46% for the shoulder position. The data input for the DNDC model were not of sufficient detail to characterize the moisture differ...

Estimation of annual nitrous oxide emissions from a transitional grassland-forest region in Saskatchewan, Canada

The increasing atmospheric N2O concentration and the imbalance in its global budget have triggered the interest in quantifying N2O fluxes from various ecosystems. This study was conducted to estimate the annual N2O emissions from a transitional grassland-forest region in Saskatchewan, Canada. The study region was stratified according to soil texture and land use types, and we selected seven landscapes (sites) to cover the range of soil texture and land use characteristics in the region. The study sites were, in turn, stratified into distinguishable spatial sampling units (i.e., footslope and shoulder complexes), which reflected the differences in soils and soil moisture regimes within a landscape. N2O emission was measured using a sealed chamber method. Our results showed that water-filled pore space (WFPS) was the variable most correlated to N2O fluxes. With this finding, we estimated the total N2O emissions by using regression equations that relate WFPS to N2O emission, and linking these regression equations with a soil moisture model for predicting WFPS. The average annual fluxes from fertilized cropland, pasture/hay land, and forest areas were 2.00, 0.04, and 0.02 kg N2O-N ha−1 yr−1, respectively. The average annual fluxes for the medium- to fine-textured and sandy-textured areas were 1.40 and 0.04 kg N2O-N ha−1 yr−1, respectively. The weighted-average annual flux for the study region is 0.95 kg N2O-N ha−1yr−1. The fertilized cropped areas covered only 47% of the regional area but contributed about 98% of the regional flux. We found that in the clay loam, cropped site, 2% and 3% of the applied fertilizer were emitted as N2O on the shoulders and footslopes, respectively.

EFFECT OF AGRICULTURE AND OF CLEAR-CUT FOREST HARVEST ON LANDSCAPE-SCALE SOIL ORGANIC CARBON STORAGE IN SASKATCHEWAN

The development of sound management approaches to reduce soil organic carbon (SOC) losses presupposes that we thoroughly understand the sources of these losses. We used a landscape-scale research design to estimate human-induced SOC losses by comparing SOC storage in undisturbed landscapes with comparable landscapes disturbed by clear-cutting of forests in the Mixedwood/Gray Luvisolic zone of central Saskatchewan and by agricultural activity in the Black soil zone. A 14.0% decrease in soil organic carbon storage in the upper 45 cm of the soil (from 57.1 Mg ha−1 in mature Mixedwood sites to 49.1 Mg ha−1 in clear-cut landscapes) occurred due to clear cutting at the research sites in the Mixedwood forest. The dominant soil type at these sites, Gray Luvisolic soils developed in glacial till, experienced a 11% loss in SOC storage; higher losses (36% loss) occurred from sandy Brunisolic inclusions in the sites. Changes in SOC storage at the research sites in the Black soil zone landscapes varied with texture an...

Landscape-scale estimates of dinitrogen fixation by Pisum sativum by nitrogen-15 natural abundance and enriched isotope dilution

Topography and slope position influence the soil and environmental factors that affect N2 fixation by legumes. The present study was conducted to (1) estimate N2 fixation by field peas in a gently rolling farm field using the natural 15N abundance and the 15N-enriched isotope dilution techniques and (2) identify soil and environmental factors that influence N2 fixation at the landscape scale. Whereas soil available water capacity, available NHinf4sup+, total crop yield, and percent N derived from N2 fixation (% Ndfa) estimated using enriched N were significantly affected by landform patterns, soil NOinf3sup-levels, seed yield, and the % Ndfa estimated using natural abundance did not follow landform patterns. The % Ndfa using natural abundance was correlated with NHinf4sup+but not with available soil water, pH, electrical conductivity, NOinf3sup-, or particle size. Estimates of the % Ndfa using enriched 15N ranged from 0 to 92.8%. The highest median value (68.6%) for % Ndfa using enriched N occurred on the divergent footslopes, with the lowest value (28.1%) on the convergent shoulders. Estimates of % Ndfa using natural abundance ranged from 13.2% to 96.9%. Smaller fluctuations during the growing season in the δ 15N of the available N pool may have resulted in less variability for % Ndfa using natural abundance compared to enriched 15N. Despite similar mean values for % Ndfa using natural abundance (44.5) and enriched 15N (49.6), no significant correlation between the two estimates was found. These results suggest that although topography may exert gross controls on N2 fixation, large variations in N2 fixation at the microsite level may preclude correlations between individual estimates and limit detection of landscape scale patterns of N2 fixation.

Soil variation within a hummocky podzolic landscape under intensive potato production

Abstract The purpose of this study was to determine if relationships between landscape position and soil properties occurred on a podzolic landscape with hummocky meso-topography, which is under intensive potato production in New Brunswick, Canada. Elevation was measured on an approximately 5 m grid spacing and used to construct a digital elevation model. Selected soil textural, morphological and chemical properties were measured on a 7×17 grid with 25 m spacing. Landform segmentation was used to divide sampling locations into eight landform elements based on profile and plan curvature, slope gradient and specific dispersal area. Orthic Humic Regosols and Orthic Sombric Brunisols were predominant in upper slope positions, whereas gleyed subgroups of the main soil orders were commonly found in the concave, lower slope positions. Lower, concave slope positions generally had lesser sand and coarse fragment contents, greater silt, organic carbon and 137 Cs contents, and greater depth to C horizon and to bedrock than upper, convex slope positions. In contrast, A horizon thickness, clay content, soil pH and soil test P showed little relationship to landscape position. Landform segmentation was effective in characterizing the spatial distribution of soil loss at the study site, as indicated based on 137 Cs content, and of soil taxa. However, the resulting landform elements had generally small differences in parameters such as SOC, sand, silt and clay content, which are commonly considered important in agricultural production. The landscape examined in this study appears to be dominated by soil loss and the closely related homogenizing effects of intensive tillage for potato production. The properties that continue to show a clear relationship to landform position are largely remnant properties (e.g., depth to bedrock or to the C horizon), whose pattern reflects a stronger hydrological and/or pedological control. The results provide further evidence that human-induced changes in soil can fundamentally alter the natural pattern of soil distribution in the landscape, even over relatively short (i.e., 100 years) time scales.

Effect of topography on nitrous oxide emissions from winter wheat fields in Central France.

We assessed nitrous oxide (N(2)O) emissions at shoulder and foot-slope positions along three sloping sites (1.6-2.1%) to identify the factors controlling the spatial variations in emissions. The three sites received same amounts of total nitrogen (N) input at 170kgNha(-1). Results showed that landscape positions had a significant, but not consistent effect on N(2)O fluxes with larger emission in the foot-slope at only one of the three sites. The effect of soil inorganic N (NH(4)(+)+NO(3)(-)) contents on N(2)O fluxes (r(2)=0.55, p<0.001) was influenced by water-filled pore space (WFPS). Soil N(2)O fluxes were related to inorganic N at WFPS>60% (r(2)=0.81, p<0.001), and NH(4)(+) contents at WFPS<60% (r(2)=0.40, p<0.01), respectively. Differences in WFPS between shoulder and foot-slope correlated linearly with differences in N(2)O fluxes (r(2)=0.45, p<0.001). We conclude that spatial variations in N(2)O emission were regulated by the influence of hydrological processes on soil aeration intensity.

Spring wheat (Triticum aestivum) yield and grain protein responses to N fertilizer in topographically defined landscape positions

A 3-yr field study was initiated in 1996 to examine the different grain yield and grain protein responses of wheat to varied N fertilizer rates in a typical glacial till landscape in Saskatchewan, Canada. Our objective was to assess the agronomic and economic feasibility of variable rate fertilizer (VRF) N application for wheat. Results suggest that spring soil water status largely determined the yield and the protein content of wheat both within different years of the study and between different landscape positions within a given year. Although grain yield was strongly related to spring soil water and was predictable on that basis, the grain yield response of wheat to fertilizer N additions was highly variable due, in part, to the dual role that N played in determining both grain yield and grain protein content. As a consequence of the unpredictable nature of the varied response of wheat to N fertilizer additions, there was little economic rationale for using VRF strategies in the 3 yr of this study. How...