Detecting, surveying, and monitoring Phytophthora ramorum in forest ecosystems.

Ellen Michaels Goheen.

Searching for Phytophthora ramorum in forest settings has been likened to looking "for a needle in a haystack." Where the pathogen currently occurs, infection patches are scattered on a rugged, highly dissected, and heavily vegetated landscape where access is often extremely difficult. P. ramorum's host list is extensive and includes members of a range of plant families representing many different growing habits. Hosts include long-lived trees that are dominant in the forest canopy, shade-tolerant trees that thrive in the midcanopy or in special habitats such as streamsides or seeps, and evergreen and deciduous shrubs, vines, and forbs that occur in the understory. Other than the outright mortality of diseased oak trees, the symptoms associated with P. ramorum infection tend to be subtle. Chlorosis may be confined to leaf tips and edges, stem bleeding dries out, and infected leaves of evergreen shrubs readily fall off and drop to the forest floor. Furthermore, the symptoms are not unique. The leaf spots, tip dieback, and bleeding associated with P. ramorum differ only subtly from symptoms caused by other leaf-inhabiting or canker-causing pathogens. Even to those familiar with its symptoms, P. ramorum is often rather "finicky." Some seasonal restrictions on sampling apply; P. ramorum is most readily isolated from its woody hosts during cool and moist periods, and samples that are allowed to dry to any degree rarely yield the pathogen in culture. Because of the relatively common occurrence of other Phytophthora species in the forest environment, including P. nemarosa that causes a bleeding canker associated with mortality in tanoaks, other diagnostic tools that are protein- or DNA-based must be unique to the species level. P. ramorum is a quarantine pest, and samples of diseased tissue must be handled carefully by trained individuals.

Despite the many challenges associated with detection, millions of acres of forests and woodlands in California and Oregon within the range of affected host species are being surveyed to determine P. ramorum 's distribution, incidence, and impact. Aerial surveys, ground-based survey designs, and image analysis are being tested and used, usually in some combination, to answer the questions of where, how much, and what is being affected. In addition, currently available information on potential hosts, likely pathways of pathogen movement, and favorable climatic conditions is being used in a risk-based analysis to plan surveys to detect P. ramorum where it may appear in the future.

Aerial Surveys

Aerial surveys to assess tree mortality have been successfully done across forested landscapes for decades. Typically, two observers fly in a fixed-wing aircraft 1,000 to 2,000 feet above ground level and map locations and attributes of dead trees. The bird's-eye view allows observers to efficiently assess large acreages and discriminate among the many signatures associated with tree species, size, and condition. Recent advances in computer mapping and Global Positioning System (GPS) technology have increased the accuracy of aerial observers to place tree locations on maps.

Systematic aerial surveys to detect oak mortality caused by P. ramorum are being done in areas of known host species in both California and Oregon. Federal and State agencies cooperate to design, conduct, and ground-truth these surveys. Observers look for recent mortality in oaks and tanoaks based on typical crown shapes and the characteristic reddish-orange to brown color of dying or dead foliage. In some cases, an additional fly-over of affected areas using rotary-blade aircraft is done to describe the sites, eliminate miscalls, and obtain an accurate location for use in on-ground navigation. Depending upon perceived risk and resources available, all or a sample of the dead oaks mapped are ground-checked by trained crews. Understory hosts occurring in the vicinity of dead oaks are also closely examined. Symptomatic tissue is taken to approved laboratories for culturing and/or DNA-based diagnostics. Database management that includes keeping up-to-date information on locations of mapped trees and survey results is an essential component of the survey.

Can aerial surveys detect P. ramorum? In trees, yes! Phytophthora ramorum-caused tanoak mortality was first detected in southwestern Oregon during a special aerial survey done in July 2001. Approximately 330,000 acres were surveyed that month. An additional and overlapping survey covering 550,000 acres was flown in Oregon in October 2001. During the two surveys 41 occurrences of groups or individual dead tanoaks were mapped and checked. Nine of these sites, ranging in size from 0.5 to 11 acres, proved positive for P. ramorum.

In 2002, over 22 million acres of host type were flown during the late spring/early summer aerial survey conducted in both California and Oregon. Approximately 500,000 acres in northern California and southwestern Oregon were surveyed again in October. Over 600 polygons of dead oaks were mapped, and ground-checking is ongoing. When the data for the 2002 aerial surveys are all in, close to 300 locations will have been ground-checked. As of late fall, six new positive P. ramorum sites (five in California and one in Oregon) had been added to the database.

Ground-based surveys and permanent plots

Detecting P. ramorum from the air is obviously nearly impossible when oak mortality is not visible or when P. ramorum -caused symptoms are restricted to leafspots and twig blights in understory species. Ground-based surveys for P. ramorum have been done in many areas within the known host type. These include roadside transect surveys designed to detect P. ramorum on leafspot and twig dieback hosts, surveys designed for early detection of P. ramorum within the host type but outside of known disease areas, and examination of permanent forest inventory plots to estimate P. ramorum -caused mortality. P.ramorum eradication treatments by the Oregon Department of Agriculture include monitoring a series of permanent plots established in the forest adjacent to treated areas. These plots are routinely visited, and vegetation is examined for symptoms of P. ramorum infection. A vast array of permanent plots have also been established to monitor P. ramorum survival, spread biology, and disease progression, among other questions.

On the ground, "the devil" is, of course, "in the details." There is a great deal of variability in the composition, stocking, and structure of the forests being surveyed. Thorough examinations of trees, shrubs, and forbs are challenging in multi-storied forest stands. How big and how many are plot questions that relate not only to statistically valid sampling schemes but also to the reality of climbing hand over hand up steep, brushy, poison oak-covered slopes where road access is limited or nonexistent.

Remote Sensing

Hundreds of miles and millions of acres of host type exist between the southern extent of P. ramorum in Monterey County, California, and its most northern occurrence in Curry County, Oregon. Interest is high in using various types of imagery and remotely sensed data to efficiently detect dead, dying, and stressed oak trees across large areas. Studies have shown that image analysis can be very useful in detecting "hot spots" of oak mortality and in quantifying even subtle changes in oak canopy structure despite limited success at identifying P. ramorum-specific mortality. In combination with other data, image analysis is expected to play a key role in low-cost, long-term monitoring of impacts associated with P. ramorum and to be a component in developing and testing risk models. (For additional information on specific image analysis projects, see http://kellylab.berkeley.edu/SODmonitoring/ )

Risk-based surveys in locations outside of affected areas in Oregon and California

Is P. ramorum present in forests in other parts of the United States? People outside of the affected areas in California and Oregon have expressed concern about the possible presence of P. ramorum in other potentially susceptible forest types. Much of this concern is based on recent inoculation studies indicating high susceptibility of some eastern oak species. Also, several understory shrub species closely related to those affected in California and Oregon occur across vast acreages of eastern oak forests.

To address this question, the USDA Forest Service Forest Health Monitoring program has taken the lead in designing survey protocols to detect P. ramorum in forest environments where it is not currently known to occur. The pilot forest survey strategy reflects the current understanding of the biology and ecology of P. ramorum , known hosts and potential hosts based on laboratory testing or taxonomic similarity, and likely pathways for its introduction. As a first step, Forest Health Monitoring produced a risk-based U.S. map identifying sampling polygons. Factors used to assign risk and develop the sampling polygons were a) presence of known host species, host genera, and closely related genera, b) locations of nurseries receiving Rhododendron spp. stock, c) length of yearly mesic/moist weather period, and d) area outside limiting temperature extremes currently associated with P. ramorum . Based on these criteria, much of the southern Appalachian and the Pacific coastal regions are currently rated high risk.

With the cooperation of several State and Federal agencies, visits to defined sampling polygons will begin in spring 2003. Trees and shrubs within plots along defined transects will be closely examined, and samples of symptomatic plants will be tested for the presence of P. ramorum at chosen laboratories. Sorting out the most likely symptoms of P. ramorum on a different group of plants will be only one of the many challenges facing the field crews conducting surveys in new areas. Future survey designs and methods will improve or change with experience gained from this pilot effort and with increased knowledge of host susceptibility, pathogen biology, and disease epidemiology. (For additional information on the National Survey and a copy of the current risk map, see http://www.na.fs.fed.us/spfo/fhm/sod/sod_natnl.pdf )

Finding, and keeping track of, the P. ramorum needle in the forest haystack is possible, but certainly not without challenges. Surveys to detect P. ramorum via air and ground will continue, and their methodologies and approaches will improve over time.