Graduate Thesis Or Dissertation
 

Projecting White Pine Blister Rust Hazard Ratings Under Climate Change in Southwestern Oregon

Público Deposited

Conteúdo disponível para baixar

Baixar PDF
https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/tq57p045j

Descriptions

Attribute NameValues
Creator
Abstract
  • Sugar pine and western white pine are widely distributed, economically valuable, and ecologically important native tree species in North America. However, white pine blister rust (WPBR), caused by a non-native fungal pathogen, Cronartium ribicola J.C. Fisch. in Rabh., has substantially affected populations of these species. Cronartium is an obligate parasite, requiring two living hosts to complete its life cycle. White pines are the primary host, and Ribes species, Castilleja, and Pedicularis spp. are alternate hosts. Once Cronartium infects trees, it damages and kills branches, stems, or entire trees by causing branch and stem cankers. Cronartium requires cool temperatures and sufficient moisture to inoculate and spread among alternate and tree hosts. Cronartium infects white pines of all size, but it is more prevalent on younger and smaller trees. Treatments to manage WPBR have included Ribes eradication, chemical spraying of Ribes, and thinning and pruning of tree hosts. However, these treatments are impractical and costly, and are not expected to eliminate WPBR. Another approach is to breed and deploy rust-resistant trees, and has been an ongoing effort for over 50 years. Finally, another option for managing WPBR is to use rust hazard ratings (RHRs). RHRs are metrics reflecting current or potential rust infection levels of sites. Specific climatic conditions, site characteristics, and tree characteristics have been used to rate the rust hazard of sites in white pine regions. For my research, I used data from Koester et al. (2018), who studied rust hazard at 265 sites in southwestern Oregon. The sites contained naturally regenerated sugar pine, naturally regenerated western white pine, rust-resistant sugar pine, and rust-resistant western white pine. I studied four rust traits, environmental variables, and tree characteristics. The rust traits consisted of percentage of trees with a stem canker (CANK%), average number of cankers per tree (NUM_CANK), average height of the highest canker (HT_CANK), and rust hazard index (RI). The environmental variables consisted of climate variables, aspect, slope, and elevation. The tree characteristics consisted of tree age, tree height, and average height growth (HT/AGE). In addition to the Koester et al data, I used the ClimateNA software program to obtain historical and future climate variables for the sites. I developed rust and tree growth random forest regression models and used other analyses to identify (1) which rust traits were best for characterizing rust hazard, (2) which environmental variables were most closely associated with rust hazard, (3) how were rust traits affected by tree characteristics, (4) how was tree growth affected by environmental variables and rust traits, (5) how will rust traits change under future climates, and (6) how will tree growth change under future climates. Based on random forest RSQ values, I concluded that CANK% and NUM_CANK were the best rust traits for characterizing rust hazard. My criteria for a good rust hazard index were that it is (1) closely associated with tree damage and death, (2) easy to measure precisely and accurately, (3) not confounded by non-rust tree variables, and (4) easy to predict from environmental variables. CANK% and NUM_CANK were relatively easy to measure precisely and accurately compared to HT_CANK and RI. Also, CANK% and NUM_CANK were more strongly associated with environmental variables, and were less confounded with non-rust tree variables. Thus, I used CANK% and NUM_CANK for further analyses. Among the environmental variables, climate variables were the most important predictors of rust hazard (i.e., based on random forest variable importance statistics). Correlations indicated that rust hazard was greater at sites with milder temperatures, sufficient moisture, and longer growing seasons. Rust hazard was also higher at sites with northern, eastern, and northeasterly aspects. Rust hazard was positively associated with tree growth and negatively associated with tree age, but had little relationship to tree height. The possible explanation is that young susceptible trees were subjected to natural selection by WPBR. Thus, older stands probably have a higher proportion of rust resistant trees. I also observed that greater tree growth was associated with harsher climates and higher rust disease, but the reasons for this were unclear. Next, I used my random forest regression models to project rust hazard and tree growth under climate change. In total, I analyzed 13 climate scenarios consisting of 12 future scenarios and one based on the 1931-1990 climate. The future climate variables were based on four Shared Socio-economic Pathways (SSPs), SSP1-26, SSP2-45, SSP3-70, and SSP5-85 and three 30-year time periods centered on 2025, 2055, and 2085. I found projected increases in spring, summer, and winter temperatures, indicating longer growing seasons in the future. In addition, summer and autumn relative humidity were projected to decrease under all SSPs. According to the projections of rust hazard, WPBR will decline, but these changes will probably be small and uncertain. Declines in WPBR may occur because warmer and drier climates inhibit rust spore germination, infection, and dispersal. In contrast, I projected a small increase in the average growth rate of the pines, probably because the models projected a longer growing season with less rust disease. The rust hazard models I developed should be particularly valuable for evaluating current rust hazards based on known climate conditions. However, many caveats make the future projections uncertain. To improve rust hazard prediction models, I recommend accounting for (1) the distributions of alternate hosts, (2) the occurrence of specific wave years, and (3) additional moisture-associated environmental variables.
Contributor
License
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Declaração de direitos
Publisher
Peer Reviewed
Language

Relações

Parents:

This work has no parents.

Em Collection:

Itens

Miniatura Título Data de carga Acesso Ações
file details CicekliZeynep2023.pdf 2023-03-24 Público Baixar