Abstract. A potentially important organizing principle in arid and semi‐arid systems is the inverse‐texture hypothesis which predicts that plant communities on coarse‐textured soils should have higher above‐ground net primary productivity (ANPP) than communities on fine‐textured soils; the reverse is predicted to occur in humid regions. Our objectives were: (1) to test predictions from the inverse‐texture hypothesis across a regional precipitation gradient, and (2) to evaluate changes in community composition and basal cover on coarse‐ and fine‐textured soils across this gradient to determine how these structural parameters may affect ANPP. Sites were located along a precipitation gradient through the Central Grassland region of the United States: mean annual precipitation ranges from 311 mm/y to 711 mm/y, whereas mean annual temperature ranges from 9 °C to 11 °C. For both coarse‐ and fine‐textured sites in 1993 and 1994, August ‐ July precipitation in the year of the study explained greater than 92% of the variability in ANPP. Soil texture did not explain a significant proportion of the variability in ANPP. However, soil texture did affect the proportion of ANPP contributed by different functional types. Forbs and shrubs made up a larger proportion of total ANPP on coarse‐ compared to fine‐textured sites. Shrubs contributed more to ANPP at the drier end of the gradient. Basal cover of live vegetation was not significantly related to precipitation and was similar for both soil textures. Our results revealed that across a regional precipitation gradient, soil texture may play a larger role in determining community composition than in determining total ANPP.

We developed a spreadsheet-based model for the use of managers, conservationists, and biologists for projecting the effects of climate change on coral reefs at local-to-regional scales. The COMBO (Coral Mortality and Bleaching Output) model calculates the impacts to coral reefs from changes in average SST and CO 2 concentrations, and from high temperature mortality (bleaching) events. The model uses a probabilistic assessment of the frequency of high temperature events under a future climate to address scientific uncertainties about potential adverse effects. COMBO offers data libraries and default factors for three selected regions (Hawai'i, Great Barrier Reef, and Caribbean), but it is structured with user-selectable parameter values and data input options, making possible modifications to reflect local conditions or to incorporate local expertise. Preliminary results from sensitivity analyses and simulation examples for Hawai'i demonstrate the relative importance of high temperature events, increased average temperature, and increased CO 2 concentration on the future status of coral reefs; illustrate significant interactions among variables; and allow comparisons of past environmental history with future predictions.

A challenge of experimental restoration is to determine the reasons why restored communities develop as they do. Divergent successions in plantings of 16 tallgrass prairie species sown in equal densities in Wisconsin (USA) revealed strong effects of vole (Microtus pennsylvanicus) herbivory on vegetation initially protected for an establishment period of over 24 months, by which time all principal species were flowering and fruiting. Half of the plots were then subjected to 48 months of vole access. An otherwise common legume (Desmodium canadense) and grass (Elymus virginicus) were all but eliminated in 48 months of exposure: combined cover under protection was 43 ± 6%, but plummeted to <1% cover when voles had access. Dicots not eaten by voles (Pycnanthemum virginianum, Rudbeckia subtomentosa) averaged 40 ± 7% cover where voles were excluded, but 68 ± 8% after 48 months of vole access. Repeated‐measures analyses of variance revealed that net decreases (D. canadense, E. virginicus) or increases (P. virginianum, R. subtomentosa) in cover masked nonlinear effects reflecting vole‐driven transient dynamics during divergent successional processes. Diversities diverged for three years after vole access; a dramatic convergence of diversities in the fourth year leaves grossly similar communities with similar numbers of species and similar levels of equity, but distinctly different species compositions. Vole‐mediated effects on vegetation resemble those likely to occur in native prairies, without the catastrophic changes in cover and standing crop caused by rodents that occurred in some previous efforts of this research program.

Habitat restoration resulting in changes in plant community composition or species dominance can affect the spatial pattern and variability of soil nutrients. Questions about how these changes in soil spatial heterogeneity develop over time at restoration sites, however, remain unaddressed. In this study, a geostatistical approach was used to quantify changes over time in the spatial heterogeneity of soil organic matter (SOM) across a 26-year chronosequence of tallgrass prairie restoration sites at FermiLab, outside of Chicago, Illinois. We used total soil N and C as an index of the quantity of SOM. We also examined changes in C:N ratio, which can influence the turnover of SOM. Specifically, the spatial structure of total N, total C, and C:N ratio in the top 10 cm of soil was quantified at a macroscale (minimum spacing of 1.5 m) and a microscale (minimum spacing of 0.2 m). The magnitude of spatial heterogeneity (MSH) was characterized as the proportion of total sample variation explained by spatially structured variation. At the macroscale, the MSH for total N decreased with time since restoration (r 2 ¼ 0.99, p < 0.001). The decrease in spatial heterogeneity over time corresponded with a significant increase in the dominance of the C 4 grasses. At the microscale, there was significant spatial structure for total N at the 4-year-old, 16-year-old, and 26-year-old sites, and significant spatial structure for total C at the 16-year-old and 26-year-old sites. These results suggest that an increase in dominance of C 4 grasses across the chronosequence is homogenizing organic matter variability at the field scale while creating fine-scale patterns associated with the spacing of vegetation. Areas of higher soil moisture were associated with higher soil N and C at the two oldest restoration sites and at the native prairie site, potentially suggesting patches of increased belowground productivity in areas of higher soil moisture. This study is one of the first to report significant changes over time in the spatial structure of organic matter in response to successional changes initiated by restoration.