Tuesday, February 1, 2022

February 2022 science summary

Pineapple the 29" tall mini horse

Hi all,

January was a bit bananas so I've only got summaries of three papers on protected areas this month (efficacy of Indigenous protected areas, recommendations to improve North American connectivity, and the importance of inventoried roadless areas in US national forests).

If you know someone who wants to sign up to receive these summaries, they can do so at http://bit.ly/sciencejon

Sze et al. 2021 compares deforestation and degradation on protected Indigenous lands, unprotected Indigenous lands, protected non-Indigenous lands, and unprotected non-Indigenous lands. Their abstract slightly misrepresents their results, which found that Indigenous lands in the tropics typically provide modest protection against deforestation and degradation, roughly similar to formal protected areas (whether Indigenous or not). The results vary by geography; in Africa unprotected Indigenous lands do even better than protected areas by most measures, but in the Americas Indigenous lands (whether protected or not) fared worse than non-Indigenous protected areas, although still better than non-Indigenous unprotected areas. In some other cases Indigenous lands seem to offer little to no improvement over unprotected non-Indigenous lands. Just comparing non-Indigenous protected areas to Indigenous protected areas, in a slight majority of cases deforestation and degradation are higher in the IPAs (but with some exceptions being similar, and degradation in Asia-Pacific being lower in IPAs); this is surprising given the other findings and makes me think that their matching process (to control for confounding variables) didn't catch everything. Another way to look at their results is that in ~90% of cases they evaluated (the 36 dark lines in Fig 2, considering both geography and data source), both protected areas and Indigenous lands (whether protected or not) experience less deforestation and degradation than unprotected non-Indigenous areas). In the remaining ~10% of cases unprotected and non-Indigenous areas have either similar levels of deforestation and degradation to protected and/or Indigenous lands, or less deforestation and degradation. Overall, the main take-away on efficacy of Indigenous lands for protection is “promising but it depends.” Their results really depend on their matching process (since without it deforestation and degradation is actually lowest in non-protected and non-Indigenous areas in a slight majority of cases). It looks like the matching should correct for confounding factors like Indigenous areas tending to be located farther from development and on lower-value lands. Most of the differences they find are pretty small. So I end up concluding that in this paper Indigenous lands are very roughly on par with protected areas, but that it’s not definitive and depends on geography.

Barnett et al. 2021 model ecological connectivity across North America to make recommendations for protected areas that best retain connectivity. The interesting part of the paper is the comparison between circuit theory and least cost approaches and how they affect recommendations. Least cost assumes species have perfect knowledge about the landscape, which is obviously untrue but over generations if individuals explore a bit on their route those routes can improve as they learn. Having the map of priorities is not terribly useful, especially since this one is based on human modification data but w/ no calibration or validation using wildlife data. The paper I wish they had written was to actually compare both modeling approaches with empirical data on wildlife movement! Essentially asking what each model gets right and wrong, and make recommendations about which approach is more useful / accurate in what context, and whether a new paradigm is needed. In my own work I’ve learned to deeply discount the value of any model which isn’t first calibrated against real world data, and then validated against other real world data not used to build the model.

Dietz et al. 2021 look at inventoried roadless areas (IRAs) in national forests in the US lower 48 states, and how important they are to vertebrate wildlife species of conservation concern (SCC - defined broadly as any of: listed under Endangered Species Act, IUCN vulnerable or worse, or NatureServe vulnerable or worse either nationally or globally, 31% of all vertebrate wildlife species). They found 57% of SCC had at least some habitat on roadless areas, and 99% of the area in IRAs provided habitat for at least one SCC. Since they're looking at about 1/3 of wildlife species, it's not shocking that intact / undeveloped forests typically provide habitat to at least SOME of those species (although as they note, since IRAs don't exist for non-forest habitats it's still impressive). The policy implications are tricky - the authors argue IRAs are good candidates for strengthening protection, but on the other hand one could argue that focusing on intact areas with less protection than IRAs would offer more benefit.

Barnett, K., & Belote, R. T. (2021). Modeling an aspirational connected network of protected areas across North America. Ecological Applications, 31(6), 1–7. https://doi.org/10.1002/eap.2387

Dietz, M. S., Barnett, K., Belote, R. T., & Aplet, G. H. (2021). The importance of U.S. national forest roadless areas for vulnerable wildlife species. Global Ecology and Conservation, 32(November), e01943. https://doi.org/10.1016/j.gecco.2021.e01943

Sze, J. S., Carrasco, L. R., Childs, D., & Edwards, D. P. (2021). Reduced deforestation and degradation in Indigenous Lands pan-tropically. Nature Sustainability, 2. https://doi.org/10.1038/s41893-021-00815-2


p.s. Pictured above is Pineapple the 29" tall mini horse. I took this photo at an event where kids in hospice (or with family members in hospice) got to hang out with horses