Coupling a spatially-explicit forest gap model with a 3-D solar routine to simulate latitudinal effects

• Published in: Ecological modelling 1996, Vol.86 (1), p.101-111
• Main Authors: Weishampel, J.F., Urban, D.L.
• Format: Article
• Language: English
• Subjects: Forest ecosystems; Latitude; Shading; Solar radiation; Spatial patterns
• ISSN: 0304-3800; 1872-7026
• DOI: 10.1016/0304-3800(94)00201-0

A multi-scaled forest model (ZELIG) which spatially embeds patch-scale processes into a larger landscape was linked with a 3-D insolation routine to simulate the effects of latitudinal variation in solar radiation on the growth and spatial patterns of idealized early successional, shade-intolerant and late successional, shade-tolerant species. At the individual tree level, average tree height increased for the shade-intolerant species with decreasing latitude as steep angled direct-beam radiation more readily penetrated narrow canopy gaps. This yielded an increase in interspecific competition and intraspecific competition among shade-intolerant trees. Consequently, though stem density dropped, basal area of shade-intolerant species increased as growth increased with decreasing latitude. Conversely, basal area for the shade-tolerant species increased with increasing latitude as stem density increased as the proportion of diffuse radiation increased. In terms of spatial pattern, the combined models produced non-random, anisotropic patterns which changed over the course of succession and were different for the shade-intolerant and tolerant species. However, differences in spatial patterns for the tropical, temperate, and boreal solar regimes were inexplicable. Shade-tolerant species consistently exhibited negative autocorrelation on adjacent grid cells (10–20 m away) during early successional stages and later stages after species replacement. Auto-correlation of shade-intolerant species shifted from negative to positive at this scale as its role changed from canopy dominant to gap colonizer. At scales > 20 m, whether or not a specific direction/distance autocorrelation value was positive or negative and significant or not significant was highly variable for both species and was attributed to stochastic properties (i.e., birth and mortality) of the model.