Seagrass stimulated by the return of the green turtle

Christopher Columbus’s journal describes how his ships had to plow through masses of sea turtles to reach the shore of Caribbean Islands. Since the 15th century, populations of the green turtle (Chelonia mydas) were nearly extirpated, primarily to feed the expanding human population. Recent conservation programs have led to a partial recovery in the Caribbean, but current green turtle populations are still a small fraction of what they were historically.

The green turtle, Chelonia mydas.

While the green turtle recovery is good news for turtles, it’s not clear how their favorite food in the Caribbean, the seagrass Thalassia testudinum feels about green turtle resurgence.  When green turtle populations were at their lowest, many lush seagrass meadows performed ecosystem services such as sequestering carbon via photosynthesis, stabilizing marine sediment  and providing important nursery habitats for commercial fisheries.  

Alexandra Gulick and her colleagues were inspired by a similar situation that has been occurring in terrestrial ecosystems.  Since around 1960, wildebeest and buffalo populations in the Serengeti have increased sharply – a result of the sharp decline (or possible elimination) of the rinderpest virus, which previously had controlled the abundance of those two large mammals.  The increase in buffalo and wildebeest populations has profoundly affected the distribution, abundance and productivity of grasses and trees, which of course impacts the entire ecosystem.  Gulick wondered whether the return of green turtles was an analogous situation, in which increases in green turtles would dramatically reduce seagrass meadows and alter ecosystem functioning.

Alexandra Gulick assisted the National Park Service and the U.S. Geological Survey with a mark recapture study of juvenile green turtles. Credit: Kristen Hart.

Gulick and her colleagues were looking for evidence of compensatory growth – increased seagrass growth in response to grazing. Green turtles use a cultivation grazing strategy, in which they select and repeatedly crop the same meadows.  Such behavior would make sense if grazed meadows compensated for grazing by producing biomass at a higher rate, or by producing leaves that were more digestible or nutritious.

A sharp boundary between a grazed and ungrazed seagrass meadow. Credit: Alexandra Gulick.

Working at the Buck Island Reef National Monument, off St. Croix, Virgin Islands, the researchers studied both grazed and ungrazed seagrass meadows, in both shallow water (3-4 meters) and deeper water (9-10 meters). They placed 129 turtle-proof exclosures over grazed and ungrazed meadows during August-October 2017 and January-February 2018. After 7-10 days they measured how much growth had occurred in both types of meadows.  

Divers set up an exclosure in a grazed meadow at Buck Island Reef National Monument. Credit: Alexandra Gulick.

The data table below shows some good evidence for compensatory growth in grazed meadows, particularly in shallow water, but also in some of the deep water meadows. Grass blades grew longer and achieved greater surface area in grazed meadows in shallow water, and also in deep water during the winter (their growth rate was slightly greater during the summer as well, but this increase was not statistically significant).  However, the seagrass in grazed meadows added much less biomass (dried weight) per day per m2.

Seagrass growth in grazed and unglazed meadows at different water depths and seasons. Mass is the increase in biomass (dry weight) per m2 per day. Statistically significant differences between grazed and unglazed meadows are boldfaced.

How can a seagrass blade have more surface area but less biomass?  There are at least two answers to this question.  First, seagrass biomass was measured on a per m2 basis, and ungrazed meadows had more blades per m2. Second, while achieving greater surface area, the seagrass blades from grazed meadows were much thinner, so when dried they weighed much less.  This is important, because putting their resources into surface area allows the seagrass blades to achieve a high photosynthetic rate, which should allow them to recover relatively quickly from sea turtle grazing.  The bottom row in the data table above is a measure of production (measured as mass growth) in relation to initial biomass (P:B).  You can see that P:B in deep water is similar in grazed vs, ungrazed meadows, while P:B in shallow meadows is substantially greater in grazed meadows. This indicates that despite continuous cropping by sea turtles, the grazed seagrass can recover quite nicely.

Gulick and her colleagues wanted to know whether the intensity of grazing might affect productivity.  They counted the number of grazed vs. ungrazed shoots, and the length of grazed vs. ungrazed blades for each sample site, and used those data to calculate grazing intensity.  The researchers then generated a model that calculated P:B in relation to grazing intensity.  The model shows that high grazing intensity increased P:B, indicating that grazing is stimulating increased leaf tissue production.

Increase in production (P:B) in relation to grazing intensity. Dashed lines indicate 95% confidence interval of the linear model.

These findings indicate that increased grazing intensity by recovering sea turtle populations is sustainable in Caribbean seagrass meadows, as seagrass growth was still stimulated at relatively high grazing intensities. Many of the meadows had been grazed continuously for at least two years, and still showed no evidence of being overly stressed by the attention that turtles had given them.  Presumably, compensatory growth by seagrass is an adaptation resulting from the co-evolution of seagrass with green turtles and other hungry herbivores. In support of this coevolution scenario, seagrass in grazed areas reduces the height of its flowers and fruits, reducing consumption of these structures by green turtles, and allowing it to achieve reproductive success.  As green turtle populations continue to recover, it is likely that seagrass meadows will be grazed more heavily, but, at least in most cases, will be able to successfully compensate for even greater grazing levels.

note: the paper that describes this research is from the journal Ecology. The reference is Gulick, A. G.,  Johnson, R. A.,  Pollock, C. G.,  Hillis‐Starr, Z.,  Bolten, A. B., and  Bjorndal, K. A..  2020. Recovery of a large herbivore changes regulation of seagrass productivity in a naturally grazed Caribbean ecosystem. Ecology 101( 12):e03180. 10.1002/ecy.3180.  Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2020 by the Ecological Society of America. All rights reserved.

Drought differentially diminishes ecosystem production

Sometimes, even the most carefully conceived experiment is thrown for a loop by Mother Nature.  Good scientists must embrace the unexpected.  Ellen Esch, David Lipson and Elsa Cleland set out to explore how plant communities responded to high, normal and low rainfall conditions.  The researchers set up rainfall manipulation plots that were covered with a clear plastic roof that would allow most light to pass through, but intercept all of the water.  They then reapplied the intercepted water, with each plot receiving either 50%, 100% or 150% of the fallen rain.  The plan was to simulate drought, normal and wet conditions. The natural world had other plans, however, as 2013-2016 were unusually dry years. Fortunately the researchers adjusted, by refocusing their question on how plant communities respond to severe drought  (50% of intercepted rainfall), moderate drought (100%) and normal rainfall (150%).


Herbaceous plant community being irrigated (notice the rainbow). Credit: Ellen Esch.

Esch and her colleagues set up their experiment at the San Diego State University Santa Margarita Ecological Reserve, which has a Mediterranean-type climate with mild, somewhat moist winters and hot dry summers.


Exotic grasses (here showing recently senesced Bromus madritensis) dominated the herbaceous sites. Credit: Ellen Esch

They wanted to know how climatic variability brought about by climate change would influence plant phenology (the timing of periodic ecological events), specifically green-up date (when plants begin turning green) and senescence date (when they turn brown and curtail photosynthesis). They expected that the native species, primarily sage-type shrubs, would be more drought-resistant than the exotic herbaceous vegetation, which was dominated by brome grass.  Climate change is predicted to increase climatic variability, which should increase the frequency and intensity of severe droughts (and also of unusually wet years).

An important measure of ecosystem functioning is its productivity – the amount of carbon taken up by an ecosystem, usually by photosynthesis.  More productive ecosystems have more energy available to feed consumers and decomposers.  More productive ecosystems also take up and store more carbon dioxide from the atmosphere, which can help reduce climate change. The researchers used a reflectance radiometer to calculate the Normalized Difference Vegetation Index (NDVI), which essentially calculates how green an area is, and is a good measure of productivity.  Esch and her colleagues hypothesized that drought would reduce overall ecosystem NDVI, but that native vegetation would be more buffered against the negative effects of drought than would the invasive exotic vegetation.


A student from a plant physiology class at San Diego State University measures NDVI. Credit: David Lipson

Each year from 2013 – 2016, the researchers set up 30 3X3 meter plots; 15 plots were dominated by exotic herbaceous species such as brome, and 15 plots had mostly native shrub species such as sage. Plots were treated the same, except for receiving either 50%, 100% or 150% of the fallen rain, which corresponded to severe drought, moderate drought and normal rainfall, respectively. Periodically, the researchers used a radiometer to measure NDVI for each plot.  They discovered that, as expected, drought reduced NDVI much more in the plots dominated by exotic herbaceous species (top graph below) than in the plots dominated by native shrubs (bottom graph).


NDVI on each measurement date for plots dominated by (top graph) exotic herbaceous species and (bottom graph) native shrub species. Red square = severe drought treatment, green circle = moderate drought, blue triangle = normal precipitation. Error bars = +/- 1 standard error.

What caused this difference in response to drought between exotic plant-dominated and native plant-dominated communities?  Mechanistically, the native shrubs have deeper roots than the exotic grasses, which may allow them to take up more water.  But how does this translate to differences in green-up date and senescence date?


A student measures stem elongation on a senescent native shrub, the black sage Salvia mellifera, near the very end of the growing season. Credit: Ellen Esch.

The researchers used two different NDVI measures to help answer this question.  Maximum NDVI is the greatest daily NDVI measure over the course of the growing season.  It is correlated with the maximum productivity of the plant community (at its greenest!).  In contrast seasonally integrated NDVI is a measure of productivity summed over the entire growing season.  Keeping those distinctions in mind, under extreme drought maximum NDVI was much lower in the exotic plots than the native plots.  But exotic plot performance increased with rainfall, so that under the wettest conditions (normal rainfall), exotic plot maximum NDVI was similar to native plot maximum NDVI (graph a below). However, when considered over the entire growing season, native plots were consistently more productive than exotic plots (graph c below).


Effect of rainfall on (a) maximum NDVI (top left), (c) seasonally integrated NDVI (top right), (b) green-up date (bottom left) and (d) senescence date (bottom right). Colors indicate dominant plot community composition (yellow = herbaceous, green = shrub) and point shape indicates growing season year (circle = 2013, square = 2014, diamond = 2015, triangle = 2016).

Phenology played an important role accounting for these differences in seasonally integrated NDVI.  At all rainfall levels, the native plant communities greened-up well before the exotic plant communities (graph b above). Exotic plants greened-up somewhat earlier as rainfall increased, while native plant green-up date was independent of rainfall. At all rainfall levels, native plots senesced about one month later than exotic plots, with increased rainfall delaying senescence in both native and exotic plant communities (graph d above).

Esch and her colleagues conclude that species composition (native shrub vs. exotic herbaceous plants) and drought both influence phenology and productivity in this important ecosystem. Climate change is predicted to increase the frequency of extreme droughts in this and other ecosystems.  Consequently, drought coupled with invasion by herbaceous species threatens to sharply reduce ecosystem productivity, which will decrease the food available for consumers and decomposers, and simultaneously reduce the amount of carbon dioxide taken up and stored by the ecosystem, thereby contributing to further climate change.

note: the paper that describes this research is from the journal Ecology. The reference is Esch, E. H.,  Lipson, D. A., and  Cleland, E. E.  2019.  Invasion and drought alter phenological sensitivity and synergistically lower ecosystem production. Ecology  100(10):e02802. 10.1002/ecy.2802. Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2019 by the Ecological Society of America. All rights reserved.