Overstory-imposed heterogeneity in solar radiation and soil moisture in a semiarid woodland

David D. Breshears, Paul M. Rich, Fairley J. Barnes, Katherine Campbell

Research output: Contribution to journalArticlepeer-review

206 Scopus citations


Degradation of semiarid ecosystems is a major environmental problem worldwide, characterized by a reduction in the ratio of herbaceous to woody plant biomass. These ecosystems can be described as a set of canopy patches comprising woody plants and the intercanopy patches that separate them, yielding an overstory with intermediate closure. Field measurements of microclimate at the scale of canopy patches, particularly for near-ground solar radiation and soil moisture, are largely lacking from both nondegraded and degraded ecosystems. We tested for relationships among spatial patterns of the over-story, near-ground solar radiation, and soil moisture in a semiarid piñon-juniper woodland in northern New Mexico that had a highly heterogeneous overstory (≈50% canopy cover) and was not degraded with respect to ground cover and erosion rates. We used measurements taken every 1 m along a 102-m transect - solar radiation indices were estimated monthly and annually using hemispherical photographs, and soil moisture was measured over 4 yr using time-domain reflectometry (TDR) - and analyzed the data using general least squares linear models that accounted for spatial autocorrelation and temporal heteroscedasticity. Time-averages of solar radiation and of soil moisture both were spatially autocorrelated at scales of up to 4 m (P < 0.05), corresponding approximately to the average lengths of both canopy and intercanopy patches and to the scale of spatial autocorrelation in the canopy/intercanopy pattern of the overstory (3 m; P < 0.05). For near-ground solar radiation, we found expected spatial variation between patches (canopy < intercanopy; P < 0.0001) and within patches for centers vs. edges (canopy center < canopy edge and intercanopy center > intercanopy edge; P < 0.0001) and for north vs. south edges (canopy north edge < canopy south edge and intercanopy south edge < intercanopy north edge; P < 0.0001). For soil moisture, canopy locations were significantly drier than intercanopy locations (P < 0.0001), and edge locations were significantly wetter than center locations both overall and within both patch types (P < 0.0001). Spatial heterogeneity in soil mositure was attributed primarily to canopy interception and drip on the basis of large differences in snow cover between canopy and intercanopy locations. Spatial autocorrelation in the residuals for soil moisture of up to 7 m was attributed to transpiration by woody plants at scales corresponding to belowground root distributions. The spatial heterogeneities in near-ground solar radiation and soil moisture are of sufficient magnitude to affect biotic processes of woody and herbaceous plants, such as growth and seedling establishment. Because land degradation problems in semiarid shrublands and woodlands appear to result from differential impacts to intercanopy vs. canopy patches, our results can be used to help design effective mitigation and remediation strategies. More generally, our results demonstrate how the physical presence of woody canopies reinforces spatial heterogeneity in microclimate and, because our site has intermediate closure of the overstory, bridge the gap along a grassland-forest continuum between related studies in relatively open savannas and in forests with nearly closed canopies.

Original languageEnglish (US)
Pages (from-to)1201-1215
Number of pages15
JournalEcological Applications
Issue number4
StatePublished - Nov 1997


  • Bouteloua gracilis
  • Canopy and intercanopy gaps
  • Juniperus monosperma
  • Land degradation
  • Microclimate
  • Overstory
  • Pinus edulis
  • Solar radiation and moisture
  • Woodland

ASJC Scopus subject areas

  • Ecology


Dive into the research topics of 'Overstory-imposed heterogeneity in solar radiation and soil moisture in a semiarid woodland'. Together they form a unique fingerprint.

Cite this