Tag Archives: plant composition

Visualizing Plant Community Change

Posted on by
16

Quick announcement: if you’re a graduate student doing research related to conservation in the Great Plains of North America, you should check out the J.E. Weaver Competitive Grant Program. We are offering up to five grants of $1,500 each to graduate students. Read more here.

About a week ago, I posted a bunch of nerdy graphs with data from some of our restored sites at the Platte River Prairies. I’ve been glad to hear from at least some of you that found those interesting. However, graphs showing how species richness or individual species occurrence changes over time are interesting, but they only show parts of the overall story. I’ve been playing with another way to better illustrate the full picture of a plant community’s species composition through time.

Photographs can help illustrate what prairies look like and how they change through time, but they only show the plants that are most visually apparent at the time of the photo. This is a 2009 photo of a restored prairie planted in the year 2000. It is one of the two sites I’m highlighting in today’s post.

What I really want to do is to step back and look at an entire plant community to examine how its composition changes through time. The concept I’m sharing today has been useful to me, but I’d love to hear feedback from others. I’d also like help naming the graphics I’ve constructed. When I started, I called them plant community signatures, but that’s doesn’t feel quite right. Help?

(For those of you who are more interested in pretty pictures than data, I did throw in a few recent photos of ice bubbles at the very end of today’s post, just for fun. If you want to see them, however, you’ll have to scroll past all my cool visual data representations first.)

The data used to create these illustrations are the same data I showed graphs of in my last data post. As a reminder, to collect the data, I plop down a 1x1m plot frame, list the plant species within it, and then repeat that over and over across a site (about 70 or more times per site). I can then calculate frequency of occurrence for each species, which is simply the percentage of those 70 or so samples each species was found in.

Once I have the data, I lay it out in a spreadsheet format to display the frequency of occurrence changes for each species through time. However, to make it easier to see patterns, I’ve colored each cell so that less frequent occurrence is light green and more frequent occurrence is dark green. I’ve also tried to group species together in a logical way. The result – at least to me – is a fascinating way to step back and study how plant composition varies over time.

I’m displaying results for two sites here. The first is a 1995 planting that I’ve collected data from annually since 2002. The second is a 2000 planting that I’ve collected data from every other year since 2003. I’m not displaying results for every species – just 60 or so of the species in each site that are abundant enough to be captured fairly regularly in 70 1x1m plots. If you want more details on these sites, you can get it from my last data-heavy post.

SITE 1 – Dahms 1995 Prairie Restoration. A 45 acre mesic site on sandy loam soils planted in spring of 1995 with approximately 150 plant species.

Here is the full graphic for the 1995 prairie restoration. It shows the frequency of occurrence between 2002 and 2023 for each featured species.

Remember to click on the title to open this post online if you’re reading it in an email. Doing that will then allow you to click on each image to see a larger, more clear version of it.

When I stare at this graphic, the first thing I notice is that there are very few examples of species that are on a distinct trajectory of increasing or decreasing frequency of occurrence. Most species simply fluctuate up and down through the years – some more dramatically than others. As expected, most of the most variable species are in the ‘annuals and biennials’ category. I think that’s great, and is evidence of ecological resilience.

Now, let’s look more closely at each of the four categories.

This graphic shows perennial grasses and sedges, including three invasive perennial grasses. Two species show a strong trend of increasing abundance. One is a native sedge (Carex brevior) and the other is Kentucky bluegrass.

Looking at the perennial grasses, it’s clear that big bluestem has been a dominant feature of this site for the duration. That’s fine, but not particularly surprising to me. What’s more interesting is the increasing abundance over time of both short-beaked sedge and Kentucky bluegrass. I’m happy about the sedge becoming more common. It’s a cool little plant. Kentucky bluegrass, of course, is less exciting because it has the potential to form monocultures and decrease plant diversity.

However, and this is really important, I’m not seeing any evidence – either here, in my other analyses, or on the ground, that Kentucky bluegrass is having a significant impact on plant diversity. Now, that could certainly change, but at least so far, it seems to be just hanging out (though it’s hanging out in a lot more places than it used to). It’s a great reminder that we should always measure the impact of invasive species, rather than just measuring the extent or abundance of the invaders themselves.

I’ll quickly highlight two other species that intrigue me. Little bluestem has become less common across the site over time, though it seems to have stabilized at about 10-20% occurrence in recent years. Simultaneously, prairie cordgrass has done the opposite, becoming more common.

Little bluestem likes drier habitats and cordgrass likes wetter. Does that mean the site has gotten wetter? I don’t think so. Instead, I think what we’re seeing is a long-term self-sorting process of a planted prairie. When the site was planted, the seed mix included way more little bluestem seed than cordgrass seed. As a result, I think the early establishment of those two species wasn’t representative of their actual adaptation to the site. Over time, I think they’ve both shifted around until they’ve settled into the microsites (and abundances) they’re best suited to.

Frequency of occurrence for perennial forbs in the 1995 planting.

I don’t see much drama happening within the perennial forbs at this site. That’s pretty reassuring, actually, given the kinds of stresses that the prairie has experienced. It’s been managed with patch-burn grazing or open gate grazing since 2002, when this data collection started. As a result, the plant community has been exposed to fire, season-long intensive grazing, and multi-year rest/recovery periods. On top of that, of course, are the wet and dry years that can also significantly affect the growth and survival of plants. It’s good to see that none of these species seems to be disappearing (which also matches up with my other analyses of our restored prairies).

The only species I see that seems to show a significant trend toward lower abundance is tall boneset. That’s not a species known to be particularly sensitive to grazing pressure, as far as I know. In fact, I think of it more as a plant that does well with the kind of lower grass dominance that comes after a bout of grazing.

My best guess is that tall boneset is a species that started out very abundant because of both lots of seed in the mix and a slow, weedy early establishment period for this planting. Over time, it might just be dropping in abundance because it’s not as competitive in this soil type when competition from its neighbors is stronger. I’m just guessing here, though. The point is that the graphic helps me identify this pattern and ask questions!

Short-lived plants and their frequency of occurrence through time.

Within the annual and biennial plants I included, it’s fascinating to see how volatile their frequency of occurrence can be from year to year. I don’t see any species with a decided trend – the species seem to just bounce up and down. What’s most intriguing to me is that all the bouncing doesn’t seem to be synchronized across species. In other words, all these short-lived plants are responding to different stimuli as they increase and decrease in their abundance.

SITE 2 – Dahms 2000 Prairie Restoration. A 69 acre mesic prairie on sandy loam soils planted over the winter between 1999 and 2000.

Now, let’s look at the second site (the one shown in the photo at the beginning of this post).

The Dahms 2000 restoration was planted with 202 species, most of which have been found at the site. Here is the full graphic for this planting, showing frequency of occurrence data for about 60 of those plant species.

I’ve only collected data from the Dahms 2000 site every other year, so there are fewer columns. Also, the site was only in its fourth growing season when I started collecting data, so we’re seeing more of the ‘early establishment’ phase of this site than we did of the 1995 planting, which was in its 8th year of growth when I first collected data there.

Frequency of occurrence for perennial grasses and sedges, including two invasive grasses.

Interestingly, there is a little more directional change within the grasses at this site than in the 1995 planting. Is that because the data includes earlier periods? Maybe, but even so, the change seems to continue past the first 8-10 years for at least some species. Big bluestem, Indiangrass, short-beaked sedge, and switchgrass all seem to be on trends of increasing frequency. Canada wildrye seems to be going in the opposite direction so it’ll be interesting to see if/where it levels off.

Both Kentucky bluegrass and smooth brome have increased in percent frequency over time. However, as in the 1995 planting, I’m not seeing any negative impacts on species diversity (yet?).

Frequency of occurrence of perennial forbs in this 2002 planting.

Within the perennial forbs, a few species at the top (stiff sunflower, yarrow, and bergamot) seem to be on an upward trend, at least before the last sampling period. You might notice that there are a lot of species – across all the categories – that dipped in frequency in 2023. I’m pretty sure that’s a consequence of thatchiness caused by a recent lack of fire and grazing, which was intentional on the part of the Platte River Prairies preserve manager (a combination of a scheduled rest period and a response to a couple dry years). I’m confident the numbers will bounce back up again over the next few years as the site gets more fire and grazing again.

Apart from those species, I think the most interesting thing about the perennial forbs is the lack of many obvious trajectories. For the most part, species seem fairly stable in their abundance through time, though some ebb and flow in interesting ways. Some species that have been labeled ‘aggressive’ in some circumstances don’t seem to be acting that way here, including Canada goldenrod and Maximilian sunflower.

Frequency of occurrence of annual and biennial plants in a 2000 prairie planting.

Just as in the first site, short-lived plants bounce around a lot in their frequency of occurrence. The variation over time is evident even though the data was only collected every two years. Once again, I don’t see much synchrony within those bounces, which tells me each species responds individually to the myriad stresses applied to the site over time. Fascinating!

So, there’s my attempt at a visual display of plant community change in our restored prairies. Does the approach seem helpful? Suggestions for improvement? Ideas for what to name the graphics?

.

And now, for those of you hoping to see pretty pictures, here are three photos of ice bubbles from a restored wetland at the Platte River Prairies last week. Enjoy!

When is a Prairie Restoration (Reconstruction) Project Successful?

Posted on by
15

This is a follow-up to last week’s post on using prairie restoration to enlarge and reconnect remnant prairies.  In this week’s post, I present a case study of a remnant sand prairie and an adjacent prairie restoration, and give thoughts about how to measure the effectiveness of that restoration project.  We’re (all of us) just getting started figuring out how to measure this kind of thing, so I’m hoping my thoughts will stimulate others to come up with their own ideas to improve upon – or contradict – mine.

Last week, I wrote about how we can improve our chances of conservation success in small isolated prairies by using prairie restoration (reconstruction) to enlarge and reconnect prairie fragments.  I even made a goofy analogy about catching falling popcorn.  At the end, I mentioned that when measuring the success of a prairie restoration – as a tool for enlarging or reconnecting remnants – we need to take a different approach than simply comparing the remnant and restored prairies to see how similar they are.  If the point of the restored prairie is to reduce the level of threat to species and natural communities inside the remnant prairie, that’s what we need to measure.

To explain what I mean, let me use a restored/remnant prairie complex along Nebraska’s Platte River as an example.  In 2000, The Nature Conservancy added several hundred acres to our Platte River Prairies through a land acquisition.  Most of the new land was cropland, but it also included 60 acres of remnant mixed-grass sand prairie with good plant diversity.  Two years later, using seed harvested from the remnant prairie and other nearby sites, we seeded 110 acres of cropland directly adjacent to the sand prairie.  The restored cropland has the same kind of hilly topography as the remnant, but also includes some low areas more appropriate for mesic tallgrass prairie.  Thus, the 162 species in our seed mixture included plant species from both mixed-grass sand prairie and mesic tallgrass prairie.

Remnant sand prairie at The Nature Conservancy's Platte River Prairies.

In June of 2010 I collected plant data from both the remnant and restored prairie (in its ninth growing season).  The data were collected by counting the plant species inside a 1m2 plot frame from 100 locations across each prairie.  Those data allowed me see the frequency of occurrence of each species (the % of plots in which each species was found).  To make the results easier for you to visualize, I’ve used a color-coding system to create what I call a plant composition signature for each prairie.  The complete comparison of the two prairies, with additional interpretation, can be seen here if you’re interested, but for this example, I’m just going to show some representative excerpts.

After the latin and common name for each species, you’ll see a column labeled “C”, which is the C-value (or coefficient of conservatism – defined by Swink and Wilhelm 1994).  If you’re not familiar with this categorization of species, a quick explanation is that lower C-value species are more opportunistic plants that can generally thrive in very disturbed environments and higher C-value species are more tied to intact native communities.  Another way to look at it is that higher C-value species are more vulnerable to habitat degradation.  All species are ranked on a scale from zero to ten (the values I’m using are specifically for Nebraska) and all exotic species get an automatic zero.

In general, the restored prairie has the same grass species as the remnant, although many are less abundant. Most of those less abundant species will spread over time as the restored prairie continues to mature. A few sedges, including sun sedge, do not establish well from seed, and we're attempting to bring them in as transplants and let them spread from there.

The main difference in "weedy" forbs between the remnant and restoration is the abundance of goldenrods in the restoration. Canada and late goldenrod were both from the seedbank, but stiff goldenrod was planted by us. At this point, I'm not concerned about the goldenrods (they don't appear to be as aggressive here as in some places) because they haven't been decreasing species diversity over time.

As with other species, I expect many of the more conservative forbs species will increase over time in the restoration.

Based on experience, I'm sure Kentucky bluegrass and smooth brome will increase over time in the restoration, but so far we've been able to manage those species to keep them from overwhelming the plant diversity in other older restorations. Apart from those two species there are no serious invaders that in the restoration that might threaten the remnant, which is good to see.

It’s easy to find differences between the remnant and restored plant communities in this example – some plant species are much more abundant in one than the other.  On the other hand, very few plant species from the remnant are missing completely from the restored prairie, and those that are less abundant are likely to increase over time.  As a prairie ecologist, I can see some obvious visual differences between the restored and remnant prairies, but most visitors to our site see the two as one large prairie.  But… Does any of this matter?  How do I decide?

First, remember that the objective of this restoration project was NOT to replicate the remnant sand prairie, but to increase the viability of the species and communities living in it.  Given that, the real questions I need to answer include the following:  Does the restored prairie increase the population size of species formerly constrained by the small remnant prairie?  Does the combination of the restored and remnant prairies provide suitable habitat for species that don’t occur in prairies the size of the remnant alone?  Does the restored prairie add to the overall resilience or ecological function of the remnant prairie?  Any questions about similarities or differences in the abundance of individual plant species need to be framed within the context of these kinds of broader questions – and tied to the specific objectives for the restoration project.  Comparisons outside of that context are relatively meaningless.

To begin evaluating the impact of the restored prairie, one first step could be to look at a few at-risk species in the remnant prairie to see if the restoration appears to benefit them.  If the remnant prairie has been harboring a small population of Franklin’s ground squirrels, for example, it’d be good to find squirrels (and their burrows) in the restored prairie as well.  If there was a rare penstemon species in the remnant (bumblebee pollinated) it’d be interesting to follow bumblebees from the plants in the remnant to see if they also visit penstemon plants in the restored prairie  – indicating that the restored prairie has facilitated growth of a genetically-interactive penstemon population.

Besides at-risk species, it would be worthwhile to search the restored prairie for the presence and/or abundance of species from other categories as well.  These categories might include:

–          Species that are representative of various types of relationships (e.g. predators and their prey, parasites/parasitoids and their hosts, insects and their larval host plants, etc.).

–          Species that have a cascading effect on other species and ecological processes (e.g. allelopathic or parasitic plants, burrowing insects/animals, etc.).

–          Species that are particularly important as food sources for a range of other species (e.g. springtails – aka Collembola, grasshoppers, “soft-bodied insects” like caterpillars and other similar larvae, etc.).

–          Area-sensitive species that may not have been able to survive in the small remnant alone but that might have a chance in the combined restored/remnant prairie (e.g. prairie chickens, badgers, and other vertebrates).

It’s also important to evaluate impacts of the restoration project on groups of species that influence ecological processes – such as pollinators and seed dispersers.  Pollinators are relatively easy to observe, and both the pollinators themselves and the resources they depend upon can be evaluated.  Ideally, of course, it’d be great to have several years of data on the species richness and abundance of pollinating insects in a small remnant prior to initiating a restoration project, followed by similar data collection after the restoration has established.    However, simply looking at whether or not purple prairie clover plants (for example) in the restored prairie are getting pollinated by the same species and numbers of pollinators as the prairie clover plants in the remnant could be very informative.  From the resource perspective, if the remnant prairie tends to lack an abundance of flowering plants at a particular time of year (late spring, for example, or early fall), measuring whether or not the restored prairie provides appropriate blooming plant species to fill that gap is very important.

Purple prairie clover being pollinated by a native bee.

There are numerous other things that could be measured, including taxonomic groups we really don’t know much about at this point.  For example, soil fauna, fungi, and obscure groups of invertebrates may very well have strong roles to play in ecological functioning of prairies, but we don’t know much about what those roles might be or how to evaluate them.  While it’s certainly important to learn more about those other taxonomic groups, our lack of knowledge shouldn’t stop us from measuring what we do know in the meantime.

The last thing to consider is whether or not a restored prairie could be actually be negatively impacting the adjacent remnant prairie or its species.  One example of this could be an invasive species that becomes established in the restored prairie – thus threatening the remnant.  A second possibility is that the restoration could function as an “ecological sink” for some species from the remnant, in which a species is drawn out of suitable habitat into attractive-looking but perilous habitat instead.  We’ve actually been testing for one possible example of this in our Platte River Prairies.  Regal fritillary larvae feed only on violets, but adults don’t lay their eggs directly on violet plants.  Our lowland remnant prairies have lots of violets, but our restored prairies have very few (so far) because we are unable to harvest large numbers of seeds.  We’re trying to make sure fritillaries aren’t laying eggs in the restorations where the larvae would be doomed to starve because of the near absence of violets.  (So far it looks like it’s not a big problem.)

As I mentioned at the beginning, we’re just starting think about how to measure the effectiveness of restored prairies as conservation tools.  Since the initial practical work of a prairie restoration project involves the establishment of a new plant community, it’s natural to assess the success of the various species we included in the seed mixture.  Unfortunately, it’s also easy to overemphasize the importance of floristic differences between a restored prairie plant community and nearby remnant prairies.  For many reasons, it’s not practical to recreate a historic prairie or replicate an existing remnant prairie.  However, it is possible to use prairie restoration to increase the viability of our remaining remnant prairies.  It is imperative to set clear objectives for this kind of restoration work, including the specific ways we want the restored prairie to help abate threats to species and communities.  Clear objectives will lead to easier decisions about how to measure success.

Many of the suggestions here are just first steps, and they and subsequent steps will require considerable resources, as well as collaboration with academic researchers.  Yes, there’s a lot to measure, but as we start to establish consistent patterns of success with some kinds of species or ecological processes, we can start focusing attention more narrowly on others.  We don’t have to test everything at once, and the most important measures at each site are those that evaluate whether or not specific objectives for that restoration project are being met.  However, it will be critical that we all share what we learn – successes and failures alike – to build up our cumulative knowledge as quickly as possible.

There are a number of examples of restoration projects where remnants have been enlarged or reconnected by restoring adjacent lands.  We should look closely at those existing sites to see if we can find evidence of success or failure (based on some of the suggested strategies above – and others).  That knowledge can guide us as we plan and implement new projects in the coming years.  It’s unlikely that we’ll be able to design restoration projects to benefit every prairie species and function, but we can certainly do a lot of good.  There’s a lot of work to be done, but I’m very optimistic about our ability to make a real difference.