A perplexing problem posed by whirling disease has been the wide variation in fish population effects from the disease. Some data sets are now sufficiently large over space and time to allow for in-depth analysis of how the parasite spreads within and across watersheds and the long-term impacts of the disease on salmonid populations. The working hypothesis of this project is that establishment and proliferation of the parasite as measured by disease risk correlates with salmonid population dynamics, and the correlations vary in space and time in relation to parasite spread and environmental risk factors such as water velocity and temperature, and level of disturbance in a watershed.  The specific objectives of the project, and results to date, are as follows.

1) Develop a technical synthesis of the “state-of-the-science” of whirling disease. This is underway. The synthesis places whirling disease into the context of other emerging aquatic infectious diseases, discussing the scientific response to disease outbreak as a potential model for other aquatic infectious disease research, and focusing on major research findings rather than exhaustively compiling all available research results.  The current draft report categorizes the major research findings to date so they can be integrated into a risk assessment approach.

2) Examine large-scale spatial and temporal patterns in parasite establishment and proliferation (i.e., infection risk) and how different risk factors correlate with infection risk. The site for this comprises eight Montana watersheds in which data on fish populations, whirling disease and stream environmental factors have been collected over many years. The investigators examined temporal and spatial patterns of disease severity in rainbow trout fry exposed in sentinel cages within and among these drainages. When sentinel cages were first deployed in Montana rivers, several were already positive for whirling disease. To model the spread of the disease, the team built a logistic regression for the response: proportion of cages showing high infection (defined as half or more of the caged fish having histology score of 3, 4, or 5) based on the predictor: time. The logistic regression is able to estimate when individual rivers became 50% infected, that is, when half the sites are predicted to have a high infection. Water temperature data were available for many of the sites, but including temperatures did not improve the fit of the model. Biotic features the team is examining include oligochaete assemblage structure, abundance, and genetic structure and infection prevalence of T. tubifex populations.  Sixteen subbasins (eight each whirling disease positive and negative) were selected to assess the relationship between disease prevalence and lineage composition and infection prevalence of T. tubifex.  Subbasins (i.e., the entire area draining into one sentinel cage location) within basins were selected such that the collections were independent of each other and paired subbasins were of similar size.  Most subbasins had 5-8 years of sentinel cage data.  Preliminary statistics indicate that subbasin area is not correlated to infection scores.

3) Examine fish population dynamics before and after the invasion of M. cerebralis. Considerable time and effort has been expended in acquiring the population datasets and integrating them into a compatible format. The team has re-analyzed population abundance data using a broad suite of competing models that allow for heterogeneity in capture probabilities to better ensure that abundance estimates are unbiased and reasonably precise. When all datasets are complete, a Before-After Control-Impact (BACI) Study Design will be used to assess population responses to whirling disease across multiple drainages. This analysis will take into account the most prominent environmental factors, as well.

Southwest regional risk assessment for whirling disease in native salmonids in arid and semi-arid lands: Arizona, Colorado, New Mexico, and Utah -
Colleen Caldwell, U.S. Geological Survey, New Mexico Cooperative Fish and Wildlife Research Unit (interim progress report)

The goal of this project is to develop a spatially-explicit risk model that will allow managers to assess the risk of whirling disease exposure and establishment within watersheds of interest. To meet this goal the investigators are characterizing the incidence and severity of M. cerebralis at a broad regional scale by compiling and analyzing environmental variables associated with M. cerebralis prevalence that include the distribution of T. tubifex, and distribution and stocking of native and non-native salmonids. Working with cooperators in Utah, New Mexico, Arizona, Colorado, Nevada and Wyoming, they have identified critical factors for mitigation or prevention of whirling disease. They’ve obtained spatially explicit occurrence, population, and age class data on native trout (Bonneville cutthroat, Colorado River cutthroat, Rio Grande cutthroat, Apache trout, and Gila trout). Other data shared by state agencies concern T. tubifex distribution, abundance and lineage (to establish areas of host overlap); waters tested for M. cerebralis; occurrence, population, and age class data of non-native trout where their range overlaps native trout populations and where disease testing has been initiated (to delineate parasite sinks and vectors for parasite transmission and location and co-occurrence of non-native trout species); watershed-specific data on organic input, beaver ponds and other landscape features that act as sediment sinks and effect on sedimentation rates; data on fish migration barriers and measures of barrier integrity; data on other potential hazards, such as public stocking sources, stocking rates, potential illicit introductions, and proximity to private ponds, roads, trails and recreation sites. Data providers were also requested to provide metadata for the submitted data. Public datasets concerning topography, hydrology, meteorology, land ownership, water quality and road networks have also been downloaded and integrated into the risk assessment framework.

As data arrive and are incorporated into the database, the conceptual model undergoes continued updating. The researchers have developed a spatially-explicit regional map depicting disease sources, supporting habitat, and native species distribution within each sub-region. They have initiated the risk ranking process and developed preliminary models for exposure, establishment, and consequences for sub-region watershed analysis. During 2008 they will calculate risk for each scenario for the sub-regions; evaluate the uncertainty and sensitivity of the models; generate testable hypotheses and validate model outcomes; and communicate the risk assessment results to stakeholders.

Myxobolus cerebralis risk to Yellowstone cutthroat trout related to variation in T. tubifex abundance and susceptibility. Billie Kerans, Montana State University (interim progress report)

The goal of this project is to develop quantitative parameters for infection risk in the tubificid host for use in an M. cerebralis risk assessment for Yellowstone cutthroat trout. This work seeks to improve understanding of T. tubifex strains and ecology and their relationships to infection of Yellowstone cutthroat trout.The specific objectives are to determine:

1. if susceptible T. tubifex strains are more prevalent and abundant where whirling disease risk is severe, 
2. whether habitat selection varies among T. tubifex strains,
3. the susceptibility of T. tubifex strains from locations with varying risk, and
4. if there are consistent patterns relating T. tubifex strains and habitat conditions, as validated in the field.

Summary of activities to date:

Objective 1: T. tubifex strains and infection risk: Infection risk was determined by infection prevalence and severity in sentinel fish in 12 tributaries of Yellowstone Lake and the Yellowstone River. 

Objective 2: T. tubifex strains and habitat characteristics:  Habitat characteristics shown to be important in other studies were measured at all sampling locations in 2005. In 2006, oligochaetes were collected from a sub-set of these sites in order to establish laboratory cultures (see objective 3), and habitat characteristics were measured concurrent with collections and at sentinel cage deployment sites. In 2007, habitat features were measured at cage sites only. Preliminary results suggest distribution of lineage III T. tubifex strains (characterized by RAPD banding patterns) may be related to specific environmental features.

Objective 3: Susceptibility of T. tubifex strains: Preliminary genetic analyses suggest at least three mitochondrial lineages may be present in Yellowstone Park. Laboratory cultures of T. tubifex have been established from oligochaetes collected from areas with different risk, where the investigators found different T. tubifex strains. Lineage markers and RAPDs suggest the laboratory cultures represent at least two mitochondrial lineages and five strains of T. tubifex populations. All clonal lines included in the experiment were lineage III T. tubifex, however, four different strains were confirmed by RAPD markers. All T. tubifex strains were susceptible to M. cerebralis, but differences in susceptibility were apparent.

Objective 4: Field validation and testing: This has not been undertaken yet. The investigators propose a change from the original workplan: to test the susceptibility of dominant strains of T. tubifex from tributaries to Yellowstone Lake and the Yellowstone River that were previously confirmed as M. cerebralis negative. They will then incorporate these data into a risk assessment for negative tributaries to Yellowstone Lake.