Holocene shifts in the assembly of the plant and animal communities implicate human impacts
S. Kathleen Lyons1, Kathryn L. Amatangelo2, Anna K. Behrensmeyer1, Antoine Bercovici1, Jessica L. Blois3, Matt Davis1,4, William A. DiMichele1, Andrew Du5, Jussi T. Eronen6, J. Tyler Faith7, Gary R. Graves8,9, Nathan Jud10,11, Conrad Labandeira1,12,13, Cindy V. Looy14, Brian McGill15, Joshua H. Miller16, David Patterson5, Silvia Pineda-Munoz17, Richard Potts18, Brett Riddle19, Rebecca Terry20, Anikó Tóth1, Werner Ulrich21, Amelia Villaseñor5, Scott Wing1, Heidi Anderson22, John Anderson22,
Donald Waller23 & Nicholas J. Gotelli24
Summarized by Altangerel Tsogtsaikhan
Background: Exploring the hitten effects and mechanisms of how ecological communities’ assemblage are crucial to making predictions of the effects of climate change. Recent work reported that the co-occurrence of species pairs are getting less and less than how it would be expected by random chance. Chronological details of how this co-occurrence structure by change is still unknown. The current study assesses the changes in plant and animal community organizations over geological time (300 million years) with quantification of 359,896 unique taxon pairs co-occurring in 80 assemblages categorizing the aggregation or segregation.
Questions: how are plant and animal communities organized, and does their structure change through time?
-Do the patterns of species segregation that characterize modern assemblages also hold in the fossil record, or is the present different?
-If whether the non-random species associations of plant and animal assemblages over the past 300 million years are dominated by segregated or aggregated species pairs?
-Why are species associations so different in fossil versus modern assemblages?
Methods: Plant datasets arranging from 307 million years ago to present including taxon presence and absence across multiple localities in different time. Also, 80 fossil data (38 mammals and 42 plants) and recent assemblage data from North America and Africa were used. Mammal dataset ranged from 21.4 ma. to present.
Each dataset was compared to a null’ assemblage generated by randomization. 39 modern communities were analyzed and compared.
Detecting of non-random species pairs: using binary presence-absence matrix for each taxon. First, the calculated C score was found.
0.0 (aggregation: maximal co-occurrence of both species)
1.0 (segregation: minimal co-occurrence of both species)
PAIRS calculates C score for each pair of species.
Next, the p-value was estimated for association with each species pair by a randomization test.
Weighted Loess regression. A Loess smoothing line was created with the stat smooth function in the R package ggplot2 version 1.0.0 (ref. 37) using default parameters.
Analysis of climate variability impacts: percentage of aggregated species pairs. Last 65 million years of assemblage dataset was used (European Project for Ice Coring in Antarctica).
Individual Evolution of Terrestrial Ecosystems Program (ETE) data set was used.
Breakpoint analysis: a maximum likelihood approach to estimate the breakpoint time (sharper decline in aggregated species in pairs).
Findings: 87-100% fossil data sets were random taxon pairs. 62 out of 80 assemblages showed significant associations that are stronger than the null model which is different than the modern mainland assemblages (most significant associations in the fossil record are aggregated in positive associations which is consistent with last 300 million years of fossil diverse assemblages include mammals, plant macrofossils, and pollen from all other continents. Around 6000 years ago (beginning of Holocene), there was a significant temporal trend towards a greater proportion of segregated species pairs were found. It is pretty similar to the modern assemblages. 64% of significant pairs were aggregated before the breakpoints which dropped to 37% after the breakpoints. A significant decrease in the percentage of positive associations were found in the Pleistocene-Holocene transition (11.700 years ago) indicating the probable human activity.
Thoughts: this is an important research for assessing macroscale pattern of community structure and their specific associations in different time scale to compare the natural versus human impact. For me it is interesting to see the time 6000. Some religious documents claim that this time as the beginning of the life formation. I wonder it has something to do with the socialized and structured human behavior. Also, I know that the Holocene climate stability is pretty unusual and I wonder if it favored the human activities and intensified the decrease in positive associations of ecological assemblages.
Were all of the analyses done with numbers of did you actually get to see what species were paired? I feel like I must know what species were BFFs and what ones were mortal enemies. It sounded like mammals were compared to mammals and plants to plants, so is it possible at all to compare mammals to plants? That would be really interesting to possibly see food web interactions.
ReplyDeleteI think that's interesting point. I just wonder if when looking at mammals and plants together there may not be many associations significant enough because even mammals feeding directly on plants typically can feed on many kinds.
DeleteI think comparing pairs of mammals directly against pairs of plants would also be interesting. However, I don't think it's possible to have one mammal and one plant make up a pair, as they have vastly different requirements and distributions.
DeleteThis is a really cool finding but it no wonder that it received some attention from critics. The findings of this paper presents controversial findings about the impact of human activity and behavior. The break in the 300 million year trend at 6,000 years is a convincing argument for agricultural impact. The greatest factor shifting animal associations would be expected to be something that cannot be imitated by a mass extinction event
ReplyDeleteNear the end of page 2 where it talks about possible drivers for human impacts lead to segregated pairs, I thought this is pretty interesting. How would increase in frequency of fire lead to segregated pairs? Maybe a sentence or two explaining the possible drivers would be helpful.
ReplyDeleteI particularly liked Fig. 1. I think it showed the trend of aggregate vs segregate pairs very nicely. Even though the time scale changes moving into the Holocene, it's clear that the slope of the proportion of aggregate pairs drops significantly into the Holocene. It is also interesting that the distribution of aggregate pairs across continents seem very similar as you go further back in time. North American pairs do seem to dominate the plot going into the Holocene.
ReplyDeleteAs I understand aggregation/segregation, these simply refer to whether two species were found together or apart more often than expected through chance. If relatively recent communities are more segregated, does this mean local community richness is lower? ...This paper made me want to create a giant map with game pieces for all the different species.
ReplyDeleteMaria´s is an interesting point. My guess is that more segregation doesn´t necessarily means lower local richness, because it can also result in the same local richness with higher higher beta diversity (??). But, as we discussed last week, the apparent trend is same richness and lower beta. No idea.
ReplyDeleteI also think it would be very interesting to see which are the pairs of aggregated and segregated species. And for which species the relation has changed. I´m sorry I will miss today's discussion.
Figure 1 looks really cool but I don't fully understand it. I know we went over something similar to this in class... but yeah. Also its always cool and sad to read about human impacts but I thought it was interesting to read about the possible drivers that are included in human impacts... jeez we are destructive.
ReplyDeleteWhy were bugs omitted? I think it was said in a previous discussion, but is it due to insects being more difficult to preserve in the fossil record? I'm not really sure what the red lines and dots represent in Fig. 2, some clarification would be helpful.
ReplyDelete