Citizen science helps predict spread of emerging infectious diseases

1. Citizen science helps predict spread of emerging infectious diseases M. Garbelotto and D. Schmidt, U.C. Berkeley R. Meentemeyer and J. Vogler, U.N.C., Raleigh & 3000-plus volunteers

2. Phytophthora ramorum: Ramorum Blight • 4+ lineages of pathogen worldwide • Origin: Vietnam/China? • Ornamental trade, worldwide • Hundreds of host species • Different severity of disease based on species

3. Phytophthora ramorum: SOD • Ramorum Blight • 4 lineages • Origin unknown • Ornamental trade, worldwide • Hundreds of species • Progressive dieback • Sudden oak death (West Coast) • 1 lineage (NA1) • Tens of native species infected • Mortality of Fagaceae and Ericaceae • Lethal girdling lesions and dieback • Introduced via infected ornamentals

4. Croucher et al. Biol Invasions 2013 Repeated introductions and not “natural” long-distance spread are responsible for extensive SOD distribution. This was determined by presence of identical genotypes at long distance and by coalescent analysis = up to 100% mortality adult tanoaks up to 70% mortality adult oaks

5. Bay- oak relationship in CA forests • CA Bay Laurel: Transmissive host • Oak: Dead-end host Oak trunks Only trunks lesion, Lesions girdle tree but are not infectious Only leaves, highly infectious

6. Oak are infected by bays less than 20 m away 3000 trees in 70 transects

7. MAJOR SURVEYING PROBLEMS • Zone of infestation is about 700 Km in length, yet less than 25% of habitat has been colonized so far • Outbreaks are not contiguous but occur in discrete patches, so there is no a clear and unique advancing front • To predict risk of infection for oaks, the scale needs to be in the tens of meters, i.e. extremely fine scale

8. SOD Blitzes • Yearly volunteer-based survey to track expansion and contraction of the pathogen’s range by sampling infectious host (bay laurel leaves) • Volunteers have to attend training and then collect over a weekend during wet spring • Local environmentalists or UC Master Gardeners organize meetings. UC Berkeley tests all samples

9. SOD Blitzes: unique features • All necessary sampling materials are provided to attendants free of charge, we do not use high tech platforms on purpose (no I Naturalist) • SOD Blitz trainings are locally organized: each may have a different angle. Volunteers can sample any location they want to • Data are confirmed by UC Lab and made public in real time over the web (SODmap.org) and using the App Sodmap mobile

10. SOD Blitz Collection Materials Flagging Tape & Pencil Instructions & Survey Collection Packet Leaf Collection Envelopes SOD Symptoms Reference Card White and Pink Data Sheets

11. Trainings: personally taught, now video has helped standardize them

12. Web-based SOD map App Sodmap mobile 170.000 observations RESULTS ARE PUBLICRed icons= SOD positive Green = SOD negative

13. App also calculates risk of oak infection at physical location of user, based on proximity to outbreaks

14. Why do citizen-scientists remain involved? Property value reduced 3–6% to 8–15%

15. Professionals vs. laypeople 2011 2012 2013 results similar to 2012 and 2011 Conclusions: volunteers attending training and collecting material are as or more likely to collect SOD-infected leaves as professionals

16. Predicted spatial distribution of disease risk through time (2008 – 2012). Predictive accuracy 0.61 0.67 0.72 0.71 n/a

17. Predictive accuracy (PA) based on generalized linear logistic regressions PAIncrease from 0.61 to 0.71 Currently the best performing predictive model for SOD !!

18. Crowdsourced data allowed for the first study on reversion on infection status in forest pathology

19. Taken together, our data suggest that : • Locally organized SOD Blitz informational sessions and training efforts were highly effective and attractive to stakeholders (program is 10 years old with high rate of return participants) • Citizen science produces useful data for understanding disease epidemiology: two “hard science” papers published • Working with citizen scientists as peers we have created the largest database in the world for a forest disease

20. 2017 results

21. Major results from 2017 Blitzes • San Luis Obispo county was negative • Highly visited tourist areas infested: risk of long distance spread • Infestations also out of forest settings, in more residential or rural areas • Hosts thought to ne marginal were badly infected and killed: e.g. manzanita

22. Conclusions • Emergent forest diseases present a major challenge in terms of monitoring and information communication • We need to join forces with stakeholders to perform research creating much needed synergy • A successful program is one that accommodates the different needs of each community, not the most high tech one • We need to share results and make recommendations in real time using novel approaches; peer reviewed process not a possibility (500 collectors 1 million users!) • DATA QUALITY AND CHANGE INDICATORS ARE ISSUES Thanks: NSF, USFS, Moore Foundation, PG&E

23. Treefaqs.org Sodblitz.org