Page 33 - Science

No insurance policy, necessarily mean annual precipitation fell
Recent data suggest that at a continental scale, those regions in Africa rich in tropical plant species (13) are also the most below 1500 mm in evergreen
sensitive to climate variability (6), implying that higher species richness does not necessarily lead to greater resilience. tropical forests in South Amer-
ica, Africa, and Southeast Asia,
possibly indicating a tipping
Vegetation sensitivity to climate variability Plant species richness
point about to be crossed. By
contrast, another study in the
Amazon rain forests found only
evidence of gradual change to
several transitional forest states
Sensitivity in response to a lengthening of
index Richness the dry season (10), rather than
100 ≥500 an abrupt change from forest to
200–500 another state.
80–200 There is also the fundamen-
50 tal question of whether switch-
30–80
ing between alternative states
12–30
0 800 0 800 is always necessarily a bad
<12
0 km km thing. Recent studies indicate
that tropical grasslands persist
disturbance apparently can also influence According to another biotic hypothesis, it in a permanent transitional state and that
the resilience of an ecosystem, and this varies is the characteristics of the component spe- the ability to switch between forest and sa-
according to vegetation type. For example, in cies (such as wood density, rooting depth, vanna in response to perturbations under-
tropical grasslands such as those in West Af- and leaf-area index) that make ecosystems pins their resilience (9). Downloaded from
rica, disturbance by fires buffers ecosystems more resilient (1). Here, some clear trends The recent studies discussed above are
from forest encroachment and thus promotes are starting to emerge. For example, Green- starting to test the many hypotheses that
grassland resilience (8). By contrast, frequent wood et al. (14) found that across forested exist to explain resilience in terrestrial eco-
disturbance by fires in tropical forest-savanna biomes, mortality rates after drought were systems. They reaffirm the complexity of
transition zones can lead to loss of resilience lower for species with greater wood den- resilience but also provide clear pointers for
in forest communities (9). sity and lower specific leaf area. A global future research and conservation. In tropical
Another abiotic attribute hypothesized to meta-analysis also identified these two ecosystems, soil type, belowground processes,
account for ecosystem resilience is soil type. characteristics as important for withstand- and rooting depth are potentially important http://science.sciencemag.org/
Again, there is some evidence to support ing drought in tropical rain forests, whereas areas of future research with direct manage-
this. In Poorter et al.’s study, high soil fertil- in tropical grasslands, plants with deeper ment applications. The factors responsible
ity had a positive influence on biomass re- roots were more resilient to drought (3). for resilience of tropical grasslands are an-
covery in Neotropical secondary forest plots other knowledge gap needing more research.
(5). Similarly, a modeling study that incor- HOW CLOSE IS A SYSTEM TO Given the importance of terrestrial ecosys-
porated remote sensing and field data pre- LOSING RESILIENCE? tem resilience to natural resource security
dicts that rain forest situated on soils with Determining which biotic and abiotic fac- and supply across the globe, research into the
low clay content will be least affected by tors contribute to resilient ecosystems is attributes underpinning it should be high on on March 1, 2018
an increase in the length of the dry season important for maintaining and enhanc- any international agenda. j
and will thus have higher resilience (10). ing them. However, when determining
REFERENCES
Belowground biotic attributes may also be conservation strategies, it is also critically
1. D. Hodgson, J. L. McDonald, D. J. Hosken, Trends Ecol. Evol.
important in determining the resilience of important to be able to identify when an 30, 503 (2015).
an ecosystem; in particular, plants that have ecosystem is about to lose its resilience and 2. B. L. Timpane-Padgham, T. Beechie, T. Klinger, PLOS ONE
root systems associated with mycorrhizal cross a threshold from a desirable to an un- 12, e0173812 (2017).
3. K. J. Willis et al., in State of the World’s Plants, 2017 Report,
fungi may have greater resilience to water desirable stable state. Several methods have
K. J. Willis, Ed. (Royal Botanic Gardens, Kew, 2017),
stress in tropical dry forests (11). been proposed to do this. For example, at pp. 42–49.
There is also a suite of biotic factors to the continental scale, Hirota et al. (15) have 4. L. E. S. Cole, S. A. Bhagwat, K. J. Willis, Nat. Commun.5,
3906 (2014).
consider. Possibly the most widely cited is shown that tropical and subtropical ecosys-
5. L. Poorter et al., Nature530, 211 (2016).
the insurance hypothesis. This suggests that tems in Africa, Australia, and South Amer- 6. A. W. R. Seddon, M. Macias-Fauria, P. R. Long, D. Benz,
more biodiverse ecosystems will be more ica switch to a savanna state when forest K. J. Willis, Nature531, 229 (2016).
7. J. Verbesselt et al., Nat. Clim. Change6, 1028 (2016).
resilient to environmental perturbations cover is less than 60%. This has direct im-
8. P. Laris, S. Dadashi, A. Jo, S. Wechsler, Plant Ecol.217, 583
because they contain a greater number of plications for the management of tropical (2016).
species available to replace functions carried forests, where deforestation is a huge issue. 9. I. Oliveras, Y. Malhi, Philos. Trans. R. Soc. B Biol. Sci.371,
out by lost species. This certainly appears to Another proposed method is to examine 20150308 (2016).
10. N. M. Levine et al., Proc. Natl. Acad. Sci. U.S.A.113, 793
be the case at the community level in some recovery rates from disturbances on the basis (2016).
ecosystems (12) but does not necessarily of the hypothesis that the closer a system is 11. K. Allen et al., Environ. Res. Lett.12, 023001 (2017).
12. M. Hisano, E. B. Searle, H. Y. H. Chen, Biol. Rev.93, 439 (2018).
recovery
hold
rate
the
slower
continental
GRAPHIC: J. YOU/SCIENCE regions with the highest tropical plant spe- will be (1). This approach appears to work in 14. S. Greenwood et al., Ecol. Lett.20, 539 (2017).
example,
to a threshold,
scale. For
the
at the
13. H. P. Linder, Front. Ecol. Evol. 10.3389/fevo.2014.00038
(2014).
cies richness in Africa (see the figure, right)
models but only seems to hold true for some
15. M. Hirota, M. Holmgren, E. H. Van Nes, M. Scheffer, Science
to climate
appear
to be
most
sensitive
(13)
example, Verbes-
ecosystems. For
terrestrial
334, 232 (2011).
perturbations
(see
(6)—the
selt et al. (7) found that recovery rates from
left)
the
figure,
opposite finding to the insurance hypothesis.
10.1126/science.aar5439
SCIENCE sciencemag.org perturbations slowed down sharply once 2 MARCH 2018 • VOL 359 ISSUE 6379 989
Published by AAAS
DA_0302Perspectives.indd 989 2/28/18 11:03 AM