Sergio Marconi

Why should we care about joint forecasting

Sergio Marconi, 11 October 2018

Joint Habitat suitability models to facilitate assisted migrations

In the perspective of dramatic changes in climate and environmental conditions, tree species will need to face adaptation and or migration strategies to avoid extinction. Despite adaptation and migration are common ecological processes in nature, they are usually slow for tree species. Strategies relying on creation of ecological corridors for natural migration may require too much time for sudden changes in climate, resulting in the extinction of sensible species. With this perspective, the core issue debated in Gray et al. (2011) and McLachlan et al. (2006) is to facilitate migration and acclimation of species by proactively transplant seeds from a particular geographic portion of its areal to a different one, previously inaccessible to that population. For example, planting seeds from a population acclimated to warm and dry environment in a colder geographical region where the same species is present. On one hand this strategy would favor acclimation of the species to changing environment, reducing the risk of its local extinction. On the other hand, the introduced genotypes could have dramatic effects on the local communities and ecosystems structure, like, for example, assume invasive behavior.

Despite the two papers seem to agree that assisted migration is a valuable conservation strategy, they seem to disagree on the scale at which planning should be done (local communities or more broad scale) and whether assisted migration should be focusing on the conservation of all species of local communities jointly (McLachlan et al, 2006) or focus on the most common and widespread species, using its specific productivity as a proxy of forest health, and avoiding generic and ad hoc management efforts (Gray et al., 2011).

Constrained assisted migration on local communities

McLachlan et al (2006) suggest that conservation strategies should be addressed with a holistic approach, aiming to preserve not only species pool and growth, but also biodiversity, habitat integrity, historic community structure, and ecosystem functions. With this perspective, it is important to pay particular attention to potential unwanted effects on local community’s disruption risk by introducing genotypes from isolated populations acclimated hundreds of kilometers away. This is particularly important because there are few or no policies to address legal introduction of threatened species and genotypes into a new environment. Similarly, there is no wide knowledge about the invasiveness potential of such species spreading to adjoining areas.

The authors seem to disagree with “more aggressive assisted migration” strategies, since they would strongly rely on ecological forecast-ability. These models predictions are fairly reliable within a short time range, while the associated uncertainty after 20 years generally becomes too high; this is a problem for the authors, since the effects of introduced species/genotypes would start to show 10-20 years after their transplantation. On the other hand, the authors seem to disagree with “avoidance of assisted migration” strategies, for they could lead to extinction of local populations anyway. The authors seem to suggest that the ideal solution is somewhere in the middle, calling it a “constrained assisted migration”. This approach would require deep knowledge of the target species-community, of their observed and expected effects on the local community. For this reason, the authors identified in (1) niche distribution modeling and (2) forecasting, (3) species interactions, (4) long distance dispersal patterns, and (5) intraspecific variation of genetic pool, the main research focus to improve our ability to understand the potential risks/opportunities of assisted migration. The authors acknowledge that time and resources are often limited, and so there are tradeoffs between ideal and applicable constrained assisted migration. They suggest that “this war will have to be fought with the army we have, not the army we want” (McLachlan et al., 2006): more focus on assisted migration studies is fundamental to reach consensus of risks and opportunities linked to assisted migration, but prompt management strategies may still be valuable and pursued when necessary.

Broad scale species-specific assisted migration plans

Gray et al., 2011 suggests that assisted migration is a potentially effective strategy, and they advocate it especially for commercially important species, within the framework of normal reforestation programs for large scale management interventions. In fact, movement of plant material is a well-established practice in North America, there is enough research on adaptation to novel environments, and niche distribution models are sufficiently reliable for such species. They suggest that just a moderate research effort is required, while “generic and ad hoc assisted migration efforts should be avoided”. The authors focus on growth and productivity related metrics to address forest ecosystem health and suggest that plans should be species-specific and focus on which seed pool to choose from. The authors acknowledge in some extent the risks involved by assisted migration, but they suggest that these risks can be mitigated by analyzing the observed biological impacts on control experiments and relying on solid scientific knowledge (robust niche distribution models). With this purpose, they propose a framework based on (1) empirical experiments (“reciprocal transplant experiments”) to detect adaptational lag; (2) remote sensing monitoring (greenness, growth and primary productivity) to address which populations of the target species are vulnerable; (3) niche distribution forecasting (process-based ecosystem modeling with predicted climate trends) to identify which species/population to choose. The paper describes the case study of Populus tremuloides in Canada. Empirical transplant experiments showed adaptation lag in a latitudinal and longitudinal gradient across the wide study area with populations adapted to warmer climates potentially best suited to migrate north, as like as those from dry environments better suited to sites expected to experience lower precipitation patterns in the future. Time series of EVI (enriched vegetation index) were used to identify sub regions most likely to be vulnerable to droughts. Finally, bioclimate envelope shifts (i.e. niche distribution models) were used to predict where to expect the species presence-absence in the short-medium-long period. The authors think that niche distribution models are reliable instruments to infer optimal seed transfer distance; confidence on which genotype to use would depend on consensus among model predictions. Despite they acknowledge the limitations in these models’ predictability, the authors suggest that such limitations should not apply for normal reforestation planning. In fact, species interactions and competition are marginal processes in reforestation, because they are controlled by spacing of plantations. Similarly, drop of general consensus of niche forecasts after 10-20 years’ time range is not important to develop successful reforestation strategies, since most vulnerable life stage of tree establishment (seedlings/sapling stage), happens within that time range.

Who makes the best case?

Despite its validity, the approach of Gray et al. (2011) presents the same defects of the aggressive assisted migration highlighted by McLachlan et al (2006). Gray et al. (2011) seem to have particular faith in the forecast ability of species distribution models. However, those models are built to predict patterns of habitat suitability at a coarser resolution than the one needed for conservation and management planning (Arenas-Castro et al., 2018). Similarly, I don’t agree with the point that uncertainty in mid-long temporal horizon (over 20 years) is not that important. As stated in McLachlan et al. (2006), the effects of dispersion of the introduced genotypes on local ecosystems have important role on whether assisted migration should be pursued or not. Those effects will likely start showing patterns after 20 years, when the established trees will have reached sexual maturity. Moreover, the approach of Gray et al. is focused on a single component of forest health: growth and productivity patterns. While forest growth and productivity largely depend on common species, rare species may be key components of other ecosystem’s functions, may have a slower migration rates and be more susceptible to extinction (Bräuer et al., 2004). Ignoring the effect of assisted migration on species interactions, could potentially threaten rare species and cause community detriment.

In light of these ideas, I think that constrained assisted migration is a more ideal approach. Such approach is more conservative but has the more holistic scope to preserve a wider spectrum of ecosystem functions and properties, better safeguarding forest health. Indeed, such approach requires deeper knowledge of the local communities in which to introduce new genotypes. Despite the ideal information required sound unfeasible for most of the cases, improvements in methods in ecology and conservation are progressively thinning these barriers. For example, McLachlan et al. (2006) suggest that one of the big issues in building successful strategies is the inability to address all species involved in a local community jointly. However, in the last decade ecologists have been focusing on developing joint distribution models that aim to predict all species conjunctly, while considering their interactions (Clark et al 2017). Despite these models are far from being perfect, they represent an opportunity to predict communities’ response both to changing environment, and relative abundance of dominant/rare species. Similarly, improvement of computational skills and the advent of big data in ecology is favoring the development of hierarchical cross-scale habitat suitability models. Such models have the advantages of address the potential effects of environment and species composition change from landscape to continental scale and predict how patterns at broader scale directly affect community assembly. Similarly, improvements in remote sensing are offering a wide spectrum of continuous estimates of habitats characters and tree functional structure (Asner et al., 2017).

To conclude, in those cases where we have to act promptly to avoid species extinctions, my general feeling is that aggressive assisted migration should be taken into consideration. In that case more continuous integrative management approach would be ideal to reduce the effect of mid-long term uncertainty (Millar et al., 2007). Management should be coupled to on-the-ground monitoring of the effects of such decisions, such that strategies can be changed promptly as unwanted effects appear on site.

References

Arenas-Castro S, Gonc ̧alves J, Alves P, Alcaraz-Segura D, Honrado JP (2018) Assessing the multi-scale predictive ability of ecosystem functional attributes for species distribution modelling. PLoS ONE 13(6): e0199292. https://doi. org/10.1371/journal.pone.0199292 Asner, Gregory P., Rainer E. Martin, D. E. Knapp, R. Tupayachi, C. B. Anderson, F. Sinca, N. R. Vaughn, and W. Llactayo. “Airborne laser-guided imaging spectroscopy to map forest trait diversity and guide conservation.” Science 355, no. 6323 (2017): 385-389.

Clark, James S., Diana Nemergut, Bijan Seyednasrollah, Phillip J. Turner, and Stacy Zhang. “Generalized joint attribute modeling for biodiversity analysis: Median‐zero, multivariate, multifarious data.” Ecological Monographs 87, no. 1 (2017): 34-56.

Clark, J.S., A.E. Gelfand, C.W. Woodall, and K. Zhu. 2014. More than the sum of the parts: forest climate response from joint species distribution models, Ecological Applications, 24:990–999.

Gray, Laura K., Tim Gylander, Michael S. Mbogga, Pei-yu Chen, and Andreas Hamann. “Assisted migration to address climate change: recommendations for aspen reforestation in western Canada.” Ecological Applications 21, no. 5 (2011): 1591-1603.

Matthies, D., Bräuer, I., Maibom, W. and Tscharntke, T., 2004. Population size and the risk of local extinction: empirical evidence from rare plants. Oikos, 105(3), pp.481-488.

McLachlan, Jason S., Jessica J. Hellmann, and Mark W. Schwartz. “A framework for debate of assisted migration in an era of climate change.” Conservation biology 21, no. 2 (2007): 297-302.

Millar, Constance I., Nathan L. Stephenson, and Scott L. Stephens. “Climate change and forests of the future: managing in the face of uncertainty.” Ecological applications17.8 (2007): 2145-2151.