Looking for Human-Nature Connections in Seasonal Wikipedia Searches

Recently, I was wrapping up some revisions on a phenology paper and to comply with the journal’s style for taxonomy, I needed to know the authority on a species of white violets that a Maine hunting guide had noted in his diaries in the mid-twentieth century. Obviously, I turned to Wikipedia.

Ecologists who study phenology (or anything!) use Wikipedia all the time, but Dr. John C. Mittermeier and his coauthors take this practice to a whole new level in their paper A season for all things: Phenological imprints in Wikipedia usage and their relevance to conservation. This study, published in PLoS Biology earlier this month, uses Wikipedia page views to trace when humans show seasonal interest in the natural world. For over 30,000 species in 245 languages —which amassed 2.33 billion pageviews between July 2015 and June 2018 — they found some strong seasonal signals linking how and when people interact with plants and animals online.

“The idea for this study happened somewhat by chance to be honest,” Dr. Mittermeier confides. “I was collecting Wikipedia pageview data on different animals as part of another study (hopefully this should be published soon!) and on a whim I decided to plot a time-series of daily views to see what it looked like.” As an ornithologist, he was drawn to migratory bird data and his whimsical time-series plot for migratory bird page views peaked near its ecological migration season. This was the prototype for a figure in the PLoS Biology paper. Mittermeier says, “this [plot] made me curious as what other plants and animals might show seasonality in their views and how widespread these patterns might be in general.”

While searching for migratory birds on Wikipedia seems categorically different from actual birding, Mittermeier and his colleagues found strong correlations between these two activities. They compared trends in Wikipedia page views to eBird records. In this analysis, eBird frequency records are like “outdoor pageviews” of bird species. “It was easy to match the eBird taxonomy to the taxonomy used by Wikipedia,” Mittermeier says, “and the way in which seasonal abundance information was structured in eBird is very accessible.”

Birders, like Wikipedia users, are surprisingly great at generating big data. Just under half of the bird species in the dataset had page view patterns correlated with seasonal eBird records. But, for species that occurred in more than one of the four language/countries (Italy, Germany, Sweden, and the U.S.), just over a third showed a significant positive relationship between eBird frequency and pageviews across multiple languages. All of the countries in this analysis are in the northern hemisphere and experiencing basically the same seasons, so I asked Mittermeier if this result indicated that some birds are more "seasonally famous" in one location? He agreed that “some species do seem to be more “seasonally famous” than others, meaning that certain species may be viewed more as seasonal indicators. This could be a result of the behavior of the species (i.e. something about their seasonality is particularly visible and obvious), some sort of cultural context (maybe the species featured in a well known book or fairy tale and had a seasonal association there, for example), or some sort of combination of both of these. Comparing how seasonal indicator species are similar or different across languages would be a great way to gain insight into what leads to a species acquiring this significance. I think this is a fascinating question and one that would be very interesting to explore further.” 

But, the paper is not limited to birds, and human interest in animal and plant Wikipedia pages is not always aligned with ecological events. Figure 2 shows a spike in shark species page views that aligns with Shark Week. There are cultural drivers to the phenology of when humans search out certain species on Wikipedia. Mittermeier shares that, “The Wild Turkey was actually the first page that I looked at in relation to cultural events. Turkeys have such a powerful association with the Thanksgiving holiday in the United States I was curious as to whether this would impact people’s online searches (it does as we show in the figure!)” When the turkey hunch worked out, Mittermeier started brainstorming other cultural or marketing events associated with plants or animals that could impact online interest. “This was right around the time that Shark Week was going on over the summer and that’s why I decided to check if that had an impact on pageviews for Great Whites.”  

While the eBird community is full of self-proclaimed bird nerds, and eBird datahas been used inpeer-reviewed papers for over a decade, the programming around Shark Week has a decidedly different relationship to science and natural history. Dr. David Shiffman, a Liber Ero Postdoctoral Fellow in Conservation Biology at Simon Fraser University, studying how information related to sharks is spread on the internet, notes, “Shark Week has a well documented problematic relationship with the truth, spreading nonsense to its massive audience that I and other scientists have to spend years correcting.” I asked him what he thought about the Wikipedia-Shark Week connection that Mittermeier and coauthors uncovered. He says, “the temporary spike in public interest in sharks that Shark Week causes is something that the marine biology community takes advantage of to spread actual facts. This paper provides further evidence that scientists wishing to engage in public outreach about their area of expertise need to know their audience, and know that there are times of year when people are more likely to be receptive to learning about that topic!” Indeed, these seasonal patterns in interest — whether for migratory birds, Thanksgiving turkeys, or sharks — can be leveraged by conservation practitioners to affect policy and outreach.

Research into the public attitudes about species, including how they rise and fall seasonally, is important. Mittermeier and his coauthors write: “Seasonal changes in human interest in plants and animals can have an important role in conservation in at least three ways: (a) by identifying species for which phenology forms a component of their “value,” (b) by helping to reveal differences or similarities in how species are valued across cultural groups, and (c) by providing temporal awareness to help maximize the effectiveness of conservation marketing campaigns.” I’ve experienced this myself in a small way: when I publish papers on spring wildflowers in the dead of winter, the press releases don’t get much traction. 

And finally, I had to address the paradox of scholarly work based on Wikipedia. I’ve TA-ed intro Biology labs and scrawled “not peer-reviewed” next to many Wikipedia-base citations in lab reports. Mittermeier laughed with me, “My mother used to teach junior high school and was always telling her students not to cite Wikipedia and now here I am using it as the source for my research.” 

Reference:

Mittermeier JC, Roll U, Matthews TJ, Grenyer R (2019) A season for all things: Phenological imprints in Wikipedia usage and their relevance to conservation. PLoS Biol 17(3): e3000146. https://doi.org/10.1371/journal.pbio.3000146 

A Little Light Reading

As the leaves fall this October and the canopies bare their skeletal limbs, there’s suddenly more light filtering across the riverside trails in Maine and I’m wearing sunglasses on runs where I used to be totally engulfed in the shade. It’s hot toddy season, pumpkin spice season, submit-your-GRFP season. When the weather finally chills we’ll get into ugly sweater season, rush-to-take-family-photos-for-a-holiday-cards season, and grading-endless-finals season. Culturally, we humans divide the year into more than just autumn-winter-spring-fall. A recent PLOS ONE paper makes the case that understory plants probably do this too.

Janice Hudson and her coauthors explored the seasonal dynamics of sunlight in a temperate deciduous forest and the ecology of the common shade-tolerant shrub, spicebush. They were inspired, in part, by a relatively obscure 1977 Ecological Monographs paper* with the unassuming title “The Distribution of Solar Radiation within a Deciduous Forest,” in which the authors, Boyd A. Hutchison and Detlef R. Matt, outline the concept of phenoseasons.  

Get ready to update your calendars — the seven phenoseasons for life under a forest canopy are: winter leafless, spring leafless, spring leafing, summer leafing, summer fully-leafed, autumnal fully-leafed, autumnal partially-leafed. I only wish that Hutchison and Matt had dined with Tolkien. Imagine the invitation: “Let’s meet for second breakfast to celebrate the end of spring leafless.” [Insert ent joke here.] Hudson was interested in how changes in light availability affected understory plants like spicebush. As Hudson explains “broadly, this study was an attempt to better understand the pre-existing conditions of the forest…[are] light conditions...a controlling factor in the distribution and presence of plant species?” The phenoseason construct hasn’t taken off in ecology and the annual cycle of subcanopy light exposure is not well understood. Hudson and her coauthors stumbled on Hutchinson and Matt while working on a literature review, but the idea of phenoseasons — now update-able with a high-tech piece of equipment called line quantum sensors — seemed ecologically intriguing. Hudson’s background is in eco-hydrology and the link between seasonal changes in light and phenology had immediate implications for her. She wanted to know “how understory plants acclimate…[and] plant contributions to nutrient and water cycling during individual phenoseasons, and yet, the literature on the subject of phenoseasons is scant.”

Hudson’s team combined a year of intense field measurements with experimentally manipulated light conditions in growth chambers to explore light intensity through the phenoseasons. At Fair Hill Natural Resource Management Area in Maryland, Hudson and her team carried a light sensor through the forest of American beech and yellow poplar trees to measure light conditions above, within, and under the spicebush canopy, compiling over 4,500 measurements in a year across 26 sites (25 in the forest, one open area just outside the forest for comparison). 

When Hudson talks about light, she talks about photosynthetically active radiation (PAR) and, for this study, subcanopy photosynthetic photon flux density (PPFD), which is a measure of PAR. Unsurprisingly, the highest PPFD values under the beech and poplar canopy occur in spring leafing — the days are growing longer, the northern hemisphere is tilted toward the Sun, the trees are still mostly bare. During summer leafing, the subcanopy PPFD values drop, and continue to decrease into summer and autumnal fully-leafed, before a slight bump for the autumnal partially-leafed phenoseason. In a nod to Hutchison and Matt, Hudson recreates their 1977 figure mapping the contours of PPFD through the year at different canopy levels with her own data. It’s the scientific equivalent of siblings re-staging family photos as adults.

But what does it mean to be a spicebush living in the light environment depicted in these figures? In general, Hudson found that there’s almost 10 times more subcanopy light available during the leafless seasons than the leafing and leafed seasons. During the leafing and leafed seasons there are high-energy sun flecks and hot spots — think of a sun-dappled forest floor — which contribute to the variability of light measurements throughout the phenoseasons. But, mostly the understory species must make proverbial hay (read: Germinate! Flower! Leaf out! Photosynthesize like crazy!) while the sun shines in the short leafing seasons. Even in the leafless seasons, the open site received much more PPFD than the subcanopy: the woody surfaces of the trees were intercepting plenty of winter and early spring light.

The spicebush plants in the field and in the growth chambers grew best under the highest PPFD conditions found in the Maryland woods. This is the light niche. In the growth chambers, plants that received higher PPFD conditions were actually less healthy, produced fewer leaves and less biomass. Hudson wrote a beautiful explanation of this when we emailed and I have to let this paragraph speak for itself:

We know that all organisms have an ecological sweet spot, but very rarely are all conditions ideal. Canopy species are "less limited" in the sense that they may experience some shading by neighbors but are primarily subject to changes in light due to latitude, season, and sky conditions. This "light intensity niche" is especially important for shade-adapted and shade-loving plant species when you consider spectral filtration (one way that plants "communicate" with each other and adapt growth direction and strategy) and temporal sequences of incident radiation at both long and short time scales (the timing and amount of light availability is crucial for physiological and biochemical processes for these species). It puts a sort of "ceiling" on the amount of light that is useful for the understory plant, whereas for canopy species there really isn't such a thing as too much light – their growth is primarily limited by the lower boundaries of light availability.

 Finally, this study’s implications for climate change research are quite interesting. In the decades while the ‘phenoseasons’ concept was languishing, research in phenology has taken off: the timing of seasonal events like leaf out and flowering are almost universally creeping earlier in response to warming temperatures. This advancing spring phenology has been definitively tracked in temperate deciduous forests like Hudson’s study site. As the climate changes, leafing phenoseasons may bite into the leafless phenoseasons. The density of the canopy may change as the species composition, size, and height of canopy trees changes. As Hudson wrote, these are the pre-existing conditions in the forest from the perspective of an understory species. We often think about species migrations and no-analog communities when we talk about the ecological effects of climate change: now I think I’ll imagine the reshuffling of the pre-existing conditions, and the interactions between biotic and abiotic factors that create the “ecological sweet spots” that we study. And now, as we enter the autumn leafless, I’ll soak up the sun on my unseasonably warm October runs. 

*This paper’s obscurity is not helped by the fact that the google scholar pdf link takes you to a 627-page annual report hot off the mimeograph with old-timey typer-writer kerning; Hutchison and Matt’s paper is buried in this report (just scroll to page 327), though much easier to find via JSTOR.

Common Gardens For All Your Climate Change Needs

A guest post from PLOS Ecology Reporting Fellows, Caitlin McDonough MacKenzie & Daniel E. Winkler, on research from the Ecological Society of America Scientific Meeting in Ft. Lauderdale, Florida, August 7-11, 2016. 

Experimental gardens are an old-school methodology. In perhaps the best known example in the 1930s and 1940s Clausen, Keck, and Hiesey transplanted Potentilla glandulosa across their range in the Sierras to explore the roles of environment and genetics played in determining growth form. Clausen, Keck, and Hiesey’s classic methodology of reciprocal transplanting has a contemporary application in climate change studies, whereby researchers relocate a plant (or seed) from its home and current climate to a transplant garden and new (and perhaps future) climate. Seven decades later, the Ecological Society of America’s 2016 Annual Meeting features experimental gardens that include species ranging from alpine forbs to douglas fir trees to a dune-loving annual—collected along latitudinal, elevation, and habitat gradients. 

Nicole Rafferty opened the Climate Change: Ranges & Phenology I session presenting her research on patterns of bumblebee visitation at the Rocky Mountain Biological Laboratory. As a part of this project, she installed a reciprocal transplant experiment with seeds from three elevations planted at 12 plots per elevation site. She wanted to test how alpine plant-pollinator relationships might change as plant communities experience new microclimates (for example, if a species is transplanted to a warmer site at a lower elevation). Unfortunately, the first year of this study coincided with a dry summer and low germination rates — as a result, in 2016 she switched to seedlings. In her 2015 seed study, the glacier lily seeds from mid-elevation had the lowest success in the transplants, suggesting that mid-elevation might be a barrier to plant migrations upslope for this species.  

Range shifts and phenological are also on the minds of researchers at the U.S. Forest Service. This time with an applied focus aimed at aiding land managers who will likely need to develop strategies to make Forest Service lands more resilient to climate change impacts. Sheel Bansal at the U.S. Forest Service’s Pacific Northwest Research Station and colleagues carried out a large-scale common garden study aptly named the Douglas-fire Seed-Source Movement trial. Their experiment used seeds from 60 sources throughout the species range in Washington, Oregon, and California and grew trees from each of the sources in 9 climatically-divergent field sites and also used artificial freeze experiments to test the impacts of changing environmental queues on Douglas fir cold hardiness and associated genetic linkages. They found strong differences in cold hardiness, with minimum winter temperatures and fall frosts as major predictors of cold hardiness based on seed source. Their results have important implications for the ability of species to shift their ranges by tracking climate envelopes, and further extend to land management efforts to maintain healthy forests experiencing future climates.

In the Great Lakes region, Elizabeth LaRue from the Emery Lab at the University of Colorado Boulder used a common garden to explore dispersal traits in American sea rocket (Cakile edentula var. lacustris). She knew that dispersal traits like pericarp, or seed wall, thickness and wet mass varied across the Cakile edentula range, but it was unclear if the variability was caused by environmental or genetic differences. Collecting seeds from across the range, and growing them together in a common garden isolated the role of genetic differences and revealed lower dispersal traits at the range edges. This data was used to inform species distribution models with different scenarios for starting dispersal genetics for Cakile edentula under climate change.

Kennedy Rubert-Nason in the Department of Entomology at the University of Wisconsin-Madison and his colleagues looked at the role of vernal freezes in determining aspen phenology and growth. They planted 6 aspen genotypes into common gardens at varying temperatures and examined a number of biological responses.  The number of days it took aspen to break bud accelerated in trees that experienced freeze-damage. Freeze-damaged trees were also stunted in their second year of growth when they experienced a freeze event during their first year. Defense compounds were also dramatically impacted, potentially indicating the negative effects of freeze events and the associated ability of the trees to defend against herbivores during their most vulnerable life stage. Their study nicely highlights the importance of the timing of environmental queues in dictating species susceptibility to a changing climate. 

Caitlin McDonough MacKenzie is a PhD candidate in the Primack Lab in the Biology Department at Boston University. She spends her field seasons in Acadia National Park, Maine studying leaf out and flowering phenology and patterns of historical species loss across plant communities. Her field methods include three ridge transects that are conveniently located adjacent to beautiful running trails and carriage roads. Away from Acadia’s granite ridges, she’s interested in underutilized sources of historical ecology data including herbarium specimens, field notebooks, photographs, and old floras; the potential for citizen science in phenology research; and the intersection of science and policy.  (Follow Caitlin on Twitter @CaitlinInMaine) 

Daniel Winkler is a PhD candidate at the University of California, Irvine and a recent National Park Service Young Leader in Climate Change. Daniel is a plant ecophysiologist interested in invasive species, evolutionary ecology, and climate change impacts on native communities in “extreme” environments. His field sites include much of the desert southwest, alpine regions of Colorado, the subalpine forests of Baja California, and the tundra of northern Japan. All of Daniel’s research focuses on climate change impacts on native systems, with an emphasis on parks and protected areas. You can follow him on Twitter @DanielEWinkler, his research on Facebook at www.facebook.com/GeoMustard/, or find more information on his website at www.winklerde.com.