Disturbance and landscape history as a reference for evaluating forest management effects at a regional scale : examples from the Coast Range of Oregon, USA

Graduation date: 2004History is an invaluable source of information to understand and evaluate\ud management influences on contemporary ecosystems and landscapes. The first two\ud chapters (Chapters 2 and 3) explored the concept of historical range of variability\ud (HRV) in landscape structure and stand structure using a stochastic fire simulation\ud model to simulate presettlement (before 1850) landscapes of the Oregon Coast Range.\ud HRV has been defined as the bounded variability of a system within constraints\ud imposed by larger-scale phenomena (e.g. climate, topography) and without significant\ud modern human influence. HRV of landscapes has been proposed as a guide for\ud biodiversity conservation in the past decade.\ud In Chapter 2, I estimated HRV of a regional landscape and evaluated the\ud similarity of current and alternative future landscapes under two land management\ud scenarios to the conditions within the HRV. The simulation results indicated that\ud historical landscapes of the region were dynamic, composed of patches of various\ud sizes and age classes ranging from 0 to> 800 years as well as numerous small\ud unburned island patches. The current landscape was outside the HRV. The landscape\ud did not return to the HRV in 100 years under either scenario largely because of lack of\ud old-growth forests and overabundance of young forests. This study showed that the\ud HRV can provide a reference condition for concrete, quantitative evaluations of\ud landscape conditions and alternative management scenarios if sufficient data exist for\ud estimating HRV. Departure from HRV can serve as an indicator of landscape\ud conditions, but results depend on scale and quantification of landscape heterogeneity. In Chapter 3, I investigated the HRV in live and dead biomass and examined\ud variability in disturbance history and forest stand development. I calculated biomass\ud as a function of disturbance history. The HRV of live and dead wood biomass\ud distributions revealed that the majority of the landscape historically contained > 500\ud Mg/ha of live wood and 50-200 Mg/ha of dead wood. The current dead wood\ud condition is outside HRV. There was a wide variation in dead wood biomass because\ud of variations in disturbance history. This study suggests that natural disturbance\ud regimes and stand development are characterized by much larger variation than is\ud typically portrayed or appreciated. The HRV approaches to evaluating landscape\ud conditions need to include both landscape and stand characteristics to better represent\ud ecological differences between managed and unmanaged landscapes.\ud In Chapter 4, I used remotely sensed data and historical vegetation data in a\ud GIS to examine changes occurred in vegetation cover since settlement in two major\ud valleys, the Coquille and Tillamook, in the region. I used existing historical\ud vegetation maps of the two valleys and collected historical vegetation data from the\ud General Land Office (GLO) survey records. I characterized current vegetation\ud conditions using an unsupervised classification of satellite images. Historically, the\ud Coquille Valley was dominated by hardwood trees and the Tillamook was by conifers.\ud Valley bottoms in both areas differed in vegetation from nearby uplands. Tree-covered\ud areas have declined substantially in both valleys as a result of agriculture and\ud development. The historical data offered reference conditions for assessment of\ud changes in biodiversity that have occurred in these unique habitats.\ud This thesis illustrates the benefit of using historical landscape information for\ud better understanding of human influence on the landscape. Historical data often have\ud many assumptions and limitations, but ecological impacts of landscape changes on\ud native biota can be better understood by comparisons with historical conditions