5. Case studies

Case study 1: Use of climate-change projections by the European and Mediterranean Plant Protection Organization in pest risk analyses of invasive alien plants

Since 2016, with the initiation of the European Union-funded project “Mitigating the Threat of Invasive Alien Plants in the European Union through Pest Risk Analysis to Support the European Union Regulation 1143/2014” (EPPO, n.d.), EPPO has considered climate change in PRAs for invasive alien species. Indeed, this is a requirement of European Union regulation 1143/2014, Article 5(d): “a thorough assessment of the risk of introduction, establishment and spread in relevant biogeographical regions in current conditions and in foreseeable climate change conditions” (European Union, 2014). Within the project, PRAs including consideration of climate change were conducted for 16 plant species: Ambrosia confertiflora, Andropogon virginicus, Cardiospermum grandiflorum, Cinnamomum camphora, Cortaderia jubata, Ehrharta calycina, Gymnocoronis spilanthoides, Hakea sericea, Humulus scandens, Hygrophila polysperma, Lespedeza cuneata, Lygodium japonicum, Pistia stratiotes, Prosopis juliflora, Salvinia molesta and Triadica sebifera.

To estimate the effect of climate change on the potential distribution of these plants, the potential distribution of each species under current and future climates was modelled using the R software package (biomod2) (Thuiller et al., 2016). Future climate conditions for the 2070s under intermediate (RCP 4.5) and higher (RCP 8.5) climate-change scenarios were obtained (the latter being considered as a worst-case scenario). The variables were obtained as averages of outputs of eight globalclimate models (BCC-CSM1-1, CCSM4, GISSE2-R, HadGEM2-AO, IPSL-CM5A-LR, MIROC-ESM, MRI-CGCM3, NorESM1-M), downscaled and calibrated against the WorldClim baseline.

Distribution maps for each scenario were produced and these were then used by expert working groups to assess the influence of climate change on entry, establishment, spread and impact. In addition, the effects of climate change that would be most relevant to the species (e.g. changes to temperature, precipitation, land use, risk of fire) were identified.

When the climate-change evaluation indicated an increased risk of entry, establishment, spread or impact in the PRA area, this was noted in the PRA as additional information. The conclusion of the PRA, however, was based on an evaluation that did not consider climate change, because of the high uncertainty related to the climate-change projections and the difficulty in capturing this uncertainty in the overall assessment.

Although the project has now ended, climate change is still included in EPPO PRAs for invasive alien plants. The models are updated along with the projected time frame: see, for example, the EPPO PRA for Solanum carolinense (gd.eppo.int/taxon/SOLCA/documents), where SSPs SSP1-2.6 and SSP3-7.0 were used to project the climate for 2041–2070.

References:

EPPO (European and Mediterranean Plant Protection Organization). n.d. LIFE project – Mitigating the threat of invasive alien plants in the European Union through pest risk analysis to support the European Union Regulation 1143/2014. Project no. LIFE PRE FR 001. In: EPPO. Paris. [Cited 26 February 2024]. www.eppo.int/RESOURCES/special_projects/ life_iap

European Union. 2014. Regulation (EU) No 1143/2014 of the European Parliament and of the Council of 22 October 2014 on the prevention and management of the introduction and spread of invasive alien species. Official Journal of the European Union L, 317: 35–55. data.europa.eu/eli/reg/2014/1143/oj (Also available at: faolex.fao.org/docs/pdf/eur140066.pdf).

Thuiller, W., Georges, D., Engler, R. &Breiner,F. 2016. biomod2: ensemble platform for species distribution modeling. R package version 3.3-7. cran.rproject.org/web/packages/biomod2/ index.html

Case study 2: Blueberry maggot establishment – use of modelling to predict expansion of pest distribution

Blueberry maggot (Rhagoletis mendax (Diptera: Tephritidae)) is an obligate fruit-parasitic fly that is native to North America and is found locally in Canada across southern Ontario and Quebec, and generally more widely distributed in New Brunswick, Nova Scotia and Prince Edward Island, as well as in the eastern part of the United States of America, south to Florida. Originally not thought to be present in Ontario and Quebec (Bush, 1966; Neilson and Wood, 1985; Vincent and Lareau, 1989), regulations were put in place to try to prevent, or at least slow, its entry. In addition, regulations were put in place in Canada (and remain in force) to prevent its transfer to British Columbia. Quebec has long been a significant producer of high-quality fresh blueberries, and trapping programmes initially ensured the continued absence of the pest from the province (Vincent and Lareau, 1989). Fresh berries for export from infested provinces to other provinces were also tested for the presence of the maggot in the berries (Dixon and Knowlton, 1994). However, by the mid-1990s, managers of commercial fields in Ontario announced the arrival of blueberry maggot, which was possibly always locally present but on alternative hosts (Smith, Gavrilovic and Smitley, 2001). At approximately the same time, the fly arrived in southern Quebec (Vincent et al., 2022; Yee et al., 2014) and regulations for Quebec were now directed to preventing entry to the Lac St. Jean area, where pesticide-free and pest-free fresh blueberries were produced in an economically important industry (Vincent et al., 2016). Over the next 10 years, the range of the pest expanded more broadly through southern Ontario, the southern shore of the St. Lawrence area of Quebec and out along the St. Lawrence River to the Gaspé region (albeit locally) and previously uninfested parts of New Brunswick, following a general northward movement. This northward movement suggests that the climate was becoming less limiting in areas where previously it was limiting.

Why the blueberry maggot was not in Lac St. Jean was a bit of a mystery. According to cold-tolerance thresholds and host availability, including the availability of wild hosts, it should have been found throughout southern Ontario and southern Quebec into the Lac St. Jean area long ago (Smith, Gavrilovic and Smitley, 2001; Vincent et al., 2016; Vincent et al., 2014). One possibility was that late spring frosts might have limited northward expansion of the fly by killing potential host fruits before oviposition could occur or before larval development was complete. Another was that unreliability of spring warmth could limit northerly expansion of the fly’s range; this could explain why massive losses of berries caused by late frost were occasionally reported in Lac Saint-Jean (C. Vincent, Agriculture and Agri-Food Canada, personal communication to M. Damus, 2010).

In 2010, the NPPO of Canada (Canadian Food Inspection Agency) was still conducting surveys and imposing regulations to prevent the entry of blueberry maggot to the Lac St. Jean region, and it was curious to know when natural arrival and establishment might occur. To attempt to answer this question, the current range of blueberry maggot in North America was used to create a bioclimatic-envelope model with the machine-learning, maximumentropy program Maxent (Phillips, Anderson and Schapire, 2006), using 1950–2000 climate norms as environmental data layers. The results were then extrapolated to climatic conditions forecast under four models of climate change (Australian CSIRO Mark 2, Canadian Centre for Climate Modelling CGCM2, Hadley Centre HadCM3 and Japanese NIES99) under two scenarios (A2a and B2a). These narrative storylines have been superseded by emissions-based scenarios, but the A2 scenario represented high human population growth and slow technological development, while the B2 scenario represented moderate population growth with more environmental protection (Nakićenović et al., 2000). In addition, a climateenvelope algorithm (BIOCLIM, Busby 1991), implemented in DIVA-GIS (Hijmans et al., 2001), was used to try to identify what the current climatic limitations to the fly’s range might be.

The current climate dataset suggested that the Lac St. Jean area was unsuitable, while the rest of southern Quebec was suitable. The future climate scenarios all suggested that Lac St. Jean would become suitable by 2020, and that the Saguenay River, which leads northwest into Lac St. Jean, would form a bridge from the infested area into the (at the time) pest free Lac St. Jean region. BIOCLIM modelling of the factors limiting establishment north of the then-infested region suggested that it was not only cold air temperature that was preventing establishment, but also the stability of the temperature. North of the St. Lawrence River, the identified most-limiting factors were temperature annual range, isothermality and temperature seasonality. The conclusion, as predicted previously by Charles Vincent (personal communication), was that it was the unreliability of seasonal weather patterns that limited northward expansion of the fly, not extreme winter cold (CFIA, 2010).

In 2018, the first adult blueberry maggot specimens were caught in the Lac St. Jean area, but no larvae were found. In 2019, more adults were caught, and by 2021, 17 out of 40 survey sites in the Lac St. Jean region were positive for blueberry maggot (Vincent et al., 2022) and the regulated area was adjusted to include the Lac St. Jean region of Quebec. While the correspondence between the predicted date of establishment and the presumed date of establishment can certainly be attributed to a great deal of luck, and modelling cannot be expected to so accurately predict arrival times, the information presented in 2010 was very helpful in managing the response of the NPPO to the finds later that decade in the Lac St. Jean area.

References:

Busby, J.R. 1991. BIOCLIM: a bioclimate analysis and prediction system. In: C.R. Margules & M.P. Austin, eds. Nature conservation – Cost effective biological surveys and data analysis, pp. 64–68. Melbourne, Australia, Commonwealth Scientific and Industrial Research Organisation.

CFIA (Canadian Food Inspection Agency). 2010. Estimate of potential distribution of Rhagoletis mendax Curran in the Lac St Jean area of Québec. Unpublished internal document. Ottawa. 11 pp.

Dixon, P.L. & Knowlton, A.D. 1994. Post-harvest recovery of Rhagoletis mendax Curran (Diptera: Tephritidae) from lowbush blueberry fruit. The Canadian Entomologist, 126(1): 121–123. doi.org/10.4039/Ent126121-1

Hijmans, R.J., Guarino, L., Cruz, M. & Rojas, E. 2001. Computer tools for spatial analysis of plant genetic resources data: 1. DIVA-GIS. Plant Genetics Resources Newsletter, 127: 15–19.

Nakićenović, N., Alcamo, J., Davis, G., de Vries, B., Fenhann, J., Gaffin, S., Gregory, K. etal. 2000. Special report on emissions scenarios. A special report of working group III of the Intergovernmental Panel on Climate Change. Cambridge, UK, Cambridge University Press. 599 pp. digitallibrary.un.org/record/466938

Neilson, W.T.A. & Wood, G.W. 1985. The blueberry maggot: distribution, economic importance, and management practices. Acta Horticulturae, 165: 171–175. doi.org/10.17660/ActaHortic.1985.165.22/p>

Phillips, S.J., Anderson, R.P. & Schapire, R.E. 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190: 231–259. doi.org/10.1016/j.ecolmodel.2005.03.026

Smith, J.J., Gavrilovic, V. & Smitley, D.R. 2001. Native Vaccinium spp. and Gaylussacia spp. infested by Rhagoletis mendax (Diptera: Tephritidae) in the Great Lakes region: a potential source of inoculum for infestation of cultivated blueberries. Journal of Economic Entomology, 94(6): 1378–1385. doi.org/10.1603/0022-0493-94.6.1378

Vincent, C. & Lareau, M.J. 1989. Update on the distribution of the blueberry maggot, Rhagoletis mendax (Diptera: Tephritidae), in Canada. Acta Horticulturae, 241: 333–337. doi.org/10.17660/ActaHortic.1989.241.57

Vincent, C., Lemoyne, P., Gaul, S.O. & Mackenzie, K. 2014. Extreme cold temperature to kill blueberry maggot (Diptera: Tephritidae) in reusable containers. Journal of Economic Entomology, 107(3): 906–909. doi.org/10.1603/EC13524

Vincent, C., Lemoyne, P., Gaul, S. & Mackenzie, K. 2016. Factors limiting the northern distribution of the blueberry maggot, Rhagoletis mendax (Diptera: Tephritidae) in Eastern Canada. European Journal of Entomology, 113(1): 143–149. doi.org/10.14411/eje.2016.018

Vincent, C., Martel, P.-O., Gagnon, J. & Morin, O. 2022. First record of blueberry maggot in Lac Saint-Jean, Quebec. Canadian Journal of Plant Science, 102(4): 940–943. doi.org/10.1139/cjps-2022-0031

Yee, W.L., Hernandez-Ortiz, V., Rull, J., Sinclair, B.J. & Neven, L.G. 2014. Status of Rhagoletis (Diptera: Tephritidae) pests in the NAPPO countries. Journal of Economic Entomology, 107(1): 11–28. doi.org/10.1603/ EC13410

Case study 3: Choice of climate-data date ranges – use of 2030 climate data in species-distribution modelling by the national plant protection organization of Canada

The first restriction that non-specialists encounter when applying specialist techniques is the availability of information accessible to persons without specialist training. Many climate-data websites are available and provide detailed climate data to the user with skills in R or other forms of programming, full training in geographical information systems, the requisite software to apply it, and so on. However, although risk assessors tend to be generalists and rely on specialists when they are available, when they are not they need to use tools that are available and “canned” – that is, ready-for-use. One excellent website with such tools is the CliMond data hub (Kriticos et al., 2012), where climate data are presented in raw format (monthly averages or totals of minimum temperature, daily maximum temperature, monthly precipitation total, daily average radiation), BIOCLIM format (35 core covariates commonly used in correlative species-distribution modelling) and CLIMEX format (location and meteorology files that are combined into a MetManager file for use in the proprietary software CLIMEX). On the CliMond website, data are provided as a baseline (historical) set, centred on 1975, and six future dates (2030, 2050, 2070, 2080, 2090 and 2100) created by two models (CSIRO and MIROC-H) under two storylines each: A1B and A2. These narrative storylines have been superseded by emissions-based scenarios (RCPs), but the A1B scenario presents a future condition after balanced emphasis on all energy sources, including fossil and renewable sources, while the A2 scenario represents high human population growth and slow technological development (Nakićenović et al., 2000). Unfortunately, the CliMond website has not been updated since 2014, but it still contains a set of relevant data for modelling potential distributions of organisms of phytosanitary concern. Because of the site’s content stability, it has particular value when non-specialists need to access ready-to-use climate data in formats required by the modelling systems most commonly applied: CLIMEX (Kriticos et al., 2015; Sutherst and Maywald 1985) and Maxent (Phillips, Dudík and Schapire, 2021).

But which data to choose? While awaiting international agreement on how to incorporate climate-change projections into risk assessment, it is nevertheless clear that it already has to be considered: projecting species-distribution models into historical climate norms (1975) no longer makes sense, and in particular for Canada, where winter cold is likely to be the major limiting factor preventing establishment of newly arrived organisms. Canada’s climate is, by virtue of its northern location and in common with other high-latitude countries, apparently warming at twice the global rate (Environment and Climate Change Canada, 2019), meaning that for these countries at least, 1975 norms no longer approach current realities. At the Plant Health Risk Assessment Unit of the Canadian NPPO (Canadian Food Inspection Agency), species-distribution models are built (trained) on data that most closely match the majority of the presence data. If the data are historical, then the baseline dataset (centred on 1975) is used and the results are projected onto another time frame. If the data are recent, and the 1975 climate norm seems too remote in time to capture the climate change that has already occurred, then both the training and projection are done with the 2030 climate dataset (all four models – both scenarios and both designs). The CliMond website unfortunately does not offer 2000, or 2010 or even 2020 data, even though these have all been observed. The closest future, most relevant dataset is then the 2030 projection, which for risk assessment of proximate near risks is the most useful and is also considered the most defensible. But the choice of 2030 is also practical: the further out one chooses, the greater the various models diverge in their predictions, and the more weight is placed on making an “accurate” guess of the future condition – that is, which climate scenario or RCP is considered most likely (Environment and Climate Change Canada, 2022). Therefore, to minimize uncertainty and match the fit-for-purpose nature of pest risk assessment, which is to address the immediate potential for harm with as little additional uncertainty as possible, a near-time future was chosen from the easily applicable and readily available data. Simply put – a pragmatic choice was made.

As the need to integrate climate change into its daily activities and planning progresses, the NPPO has added specialists to the science staff of its plant-health programme. It is expected that they will soon conduct species-distribution modelling for risk-assessment purposes in a way that the generalist risk assessors could not, but for the time being, the current approach seems to have been successful.

References:

Environment and Climate Change Canada. 2019. Canada’s climate is warming twice as fast as global average. In: Government of Canada, 2 April 2019. Ottawa. [Cited 16 December 2022]. www.canada.ca/en/environment-climatechange/news/2019/04/canadas-climate-iswarming-twice-as-fast-as-global-average.html

Environment and Climate Change Canada. 2022. Understanding future projections – Uncertainty in climate projections. Learning zone topic 3. Canada, Climatedata.ca. [Cited 16 December 2022] climatedata.ca/resource/ uncertainty-in-climate-projections/

Kriticos, D.J., Maywald, G.F., Yonow, T., Zurcher, E.J., Herrmann, N.I. & Sutherst, R.W. 2015. Climex version 4 – Exploring the effects of climate on plants, animals and diseases. Canberra, CSIRO. ix + 172 pp.

Kriticos, D.J., Webber, B.L., Leriche, A., Ota, N., Macadam, I., Bathols, J. & Scott, J.K. 2012. CliMond: global high-resolution historical and future scenario climate surfaces for bioclimatic modelling. Methods in Ecology and Evolution, 3: 53–64. doi.org/10.1111/j.2041-210X.2011.00134.x

Nakićenović, N., Alcamo, J., Davis, G., de Vries, B., Fenhann, J., Gaffin, S., Gregory, K. et al. 2000. Special report on emissions scenarios. A special report of working group III of the Intergovernmental Panel on Climate Change. Cambridge, UK, Cambridge University Press. 599 pp. digitallibrary.un.org/record/466938

Phillips, S.J., Dudík, M. & Schapire, R.E. 2021. Maxent software for modeling species niches and distributions (Version 3.4.1). [Cited 16 December 2022] biodiversityinformatics.amnh.org/ open_source/maxent/

Sutherst, R.W. & Maywald, G.E. 1985. A computerized system for matching climates in ecology. Agriculture, Ecosystems and Environment, 13: 281–299. doi. org/10.1016/0167-8809(85)90016-7