HISTORICAL RANGE OF VARIABILITY IN LANDSCAPE STRUCTURE: A SIMULATION STUDY IN OREGON, USA

Abstract

2 Pacific Northwest Research Station, USDA Forest Service, Corvallis, Oregon 97331 USA Abstract. We estimated the historical range of variability (HRV) of forest landscape structure under natural disturbance regimes at the scale of a physiographic province (Oregon Coast Range, 2 million ha) and evaluated the similarity to HRV of current and future landscapes under alternative management scenarios. We used a stochastic fire simulation model to simulate presettlement landscapes and quantified the HRV of landscape structure using multivariate analysis of landscape metrics. We examined two alternative policy sce- narios simulated by two spatially explicit simulation models: (1) current management pol- icies for 100 years into the future and (2) the wildfire scenario with no active management until it reached the HRV. The simulation results indicated that historical landscapes of the province were dynamic, composed of patches of various sizes and age classes ranging from 0 to .800 years including numerous, small, unburned forest islands. The current landscape was outside the HRV. The landscape did not return to the HRV in the 100 years under either scenario, largely because of lack of old-growth forests and the abundance of young forests. Under the current policy scenario, development of landscape structure was limited by the spatial arrangement of different ownerships and the highly contrasting management regimes among ownerships. As a result, the vegetation pattern after 100 years reflected the ownership pattern. Sur- prisingly, the wildfire scenario initially moved the landscape away from the HRV during the first 100 years, after which it moved toward the HRV, but it required many more centuries to reach it. Extensive forest management and human-caused fires in the 20th century have left legacies on the landscape that could take centuries to be obliterated by wildfire. Departure from the HRV can serve as an indicator of landscape conditions, but results depend on scale and quantification of landscape heterogeneity. The direct application of the concept of HRV to forest policy and management in large landscapes is often limited since not all ownerships may have ecological goals and future climate change is anticipated. Natural disturbance-based management at large scales would not show the projected effects on landscape structure within a typical policy time frame in highly managed landscapes.