Summary of a massive new analysis of monarch abundance data, which shows STABLE breeding populations
Hello blog readers,
The title of this blog post is not a typo, and you read it correctly. Today I'm going to tell you about a brand new monarch study that you simply must read. It is a recently-completed manuscript that I just uploaded to a preprint server, and which I also just now submitted to a scientific journal for peer review. That's right, it has not been peer-reviewed yet. My coauthors and I felt that it is important that this paper be made public prior to peer-review, because of the significance of our findings. So, here is the link to it - completely public and downloadable. And even better, I'm going to tell you all about this new project in this blog post. So strap in...it's going to get real.
OK, let me back up just a bit. If this all sounds familiar to long-time readers of this blog it's because it sort of is. Recall that in late 2020 I had blogged about a related project (link here). At that time I had conducted a year-long undertaking to write a review paper that examined all publicly-available long-term datasets on monarch abundance. This was all happening as the USFWS was deliberating on whether to list the monarchs as "Threatened" under the Endangered Species Act. At the time I had written a manuscript, uploaded it to a preprint server, and submitted it to a peer-reviewed journal. Well, the reviewers of the paper did not give it a favorable review, not because it was wrong, or flawed, or anything like that, but they felt it lacked "statistical rigor". See, I had originally intended the paper to be a very wordy and lengthy summary or review of already-published data, since initially it was never my intention to go hog-wild with stats and graphs and figures to make my point. But, this is clearly what the reviewers wanted. They were simply not convinced of the conclusion until they had "seen the data", and until the data had been more rigorously analyzed.
So...long-story short, here it is almost a year later, and I have a brand new version completed. And you guessed it, this version has oodles of stats and figures, and even more data than last time. This was all thanks to a fortuitous recent collaboration I've stuck up with some great scientists in the entomology department here at UGA, Mike Crossley and Bill Snyder, and, with another great scientist, Matt Moran, from Hendrix College in Arkansas. All of us together have combined forces to turn my former review paper into a statistical powerhouse of a research paper.
Here's the crazy part - the end result is the same. The main conclusion from this new study, after "rigorous" statistical analyses of a boatload of data, is that the abundance of monarchs during the breeding season in North America has not declined. That was my conclusion before!
Let's get into the details now, because I bet everyone reading this will also not be convinced until they see the data too. For this study we examined the long-term trends of monarch abundance from 15 different monitoring projects conducted in North America. Most of these were the same data sources that I had included in the prior version, and most of these were published or publicly-accessible data. These datasets included 2 from the spring period, 7 that tracked monarchs during the summer, and 6 that tracked monarchs during the fall migration. We did NOT examine long-term trends in winter colony size in the eastern or western subpopulations. This is one clear difference over the prior version. We figured that there are indeed clear declines happening at the winter colonies and these are not in dispute, so we didn't bother to include them here. What IS in dispute, is whether these winter declines matter.
That's not all, in this new version we also included the raw data on monarch abundance from the long-running North American Butterfly Association (NABA), and here, we used all data, from all states and provinces, over the entire time frame of the program - that's right - every scrap of data, including sites in the Midwest, Northeast, Southeast, and even in the western US. In fact we even have Jeff Glassberg, the former president of NABA, as an author on the paper, because these data ended up being such a big part of the study. This was also a big improvement over the prior iteration, where these site-level data were not included. Collectively, there were 134,684 records of monarchs from 341 sites across North America!
As you can see, we had a virtual mountain of data and records of monarchs spanning many years to comb through. Collectively we had more data than any other project to date. The details of each dataset are fully-available in our paper, and including in our supplemental file. The thing that I appreciate most about these data is that they come from different organizations, from different places, and using different methodologies. Some were butterfly atlas censuses, some were walking transects, some were counts of monarchs in museums, and one dataset was of the relative amount of monarchs that comes from the Midwest each year, based on isotopic analyses of wing tissue! We figured that to truly understand if the monarch population is going up or down, you need to examine every scrap of information possible. And, if each of these projects shows something similar, despite their methodological differences, then this tells us the results are real.
I'm next going to show you the main figures from the paper, and get into the results.
First, we examined the long-term trends from each of the 15 different studies and programs, and each trend is shown in this graph below. They are arranged in order of season. Don't worry about the units on the Y axis, the graphs have been formatted so that they are all on the same scale. The only thing to look for here is the direction of the line, and whether it is statistically significant (where the p is less than 0.05). In other words, if a line shows a hint of a trend upwards or down, but it is not significant, then it is meaningless. Also note, for the statistically-inclined scientists reading this, to analyze these trends we used "generalized least squares regression and included a first-order autocorrelation term to account for temporal autocorrelation in residuals."
Looking at these graphs, you will note that there are a few datasets that show a statistically significant decline, such as those collected during the spring recolonization. This makes sense, since we know that the winter colonies have been declining over time, and therefore so too, should the number of spring migrants. But, then when you look at the data collected during the summer, there is one that shows a decline, one shows an increase, and 5 show no decline. Similarly, of the fall monitoring projects, one shows a decline, one shows an increase, and 4 show no change. So from this part of our paper, we concluded that while there are a few datasets that show declines, most show increases or no change.
Next, we examined the mountain of data from NABA, to see if we could find any evidence of a change in monarch abundance over time. We (really, Mike) compiled the sightings of monarchs at each site (all 341) across North America going back to the mid 1990s up to 2018, and conducted a very, very detailed statistical analysis of these data. This analysis also included data on the amount of cropland present at each site, the amount of roundup used at the site, the climate at each location, and even the road density at each location! In other words, we controlled for any possible factor that might explain the variation in monarch abundance, so that we could tease out whether there was an increase or decrease in monarch numbers.
Before I get into the nitty gritty here, I want to point something out here about the NABA survey design, because here is where some people get confused. The "sites" where people track butterfly numbers each year are not just a backyard, or a park. They are actually very large swaths of landscapes where people drive around, walk, and watch for butterflies, and they do this for an entire day in the middle of each summer. The area of a NABA site is a circle with a diameter of 15 miles. This is equivalent to 175 square miles! That's the equivalent of a large county. Below is a diagram of two typical NABA circles, to give you an idea of how large they are, and, what their landscapes are like.
We chose these two sites as examples to show people that some of these sites are located in the heart of the Midwest, where there is nothing but corn and soy within the circle. This is a very important point, because of the widely-held belief that the monarchs have declined due to the loss of milkweed in agricultural fields over the last 25 years. We figured that if this was the case, then those sites located in the Midwest, and with lots of corn and soy around them, should have seen a decrease in monarch abundance.
Here is what we found from the NABA data, and I'll show you our main map of all of the NABA sites first.
This map shows each NABA site across North America, and they are color-coded so that sites in blue saw an increase in monarchs over time, and orangey sites show where there was a decrease over time. The darkest colors show where the trend was statistically significant (i.e. real). You can go ahead and try to count them up and see if the orange dots outnumber the blues, but I can tell you that they don't. In fact, we calculated that 88% of the dots show no real trend at all. Then, 8% of the sites saw a significant increase, while just 4% of sites saw a real decrease in monarch abundance. We concluded that while there were a few places where monarchs were declining, and there were some where they were increasing, most showed no change over time. Recall that this is the same conclusion we reached from the first set of analyses above.
Now, let's talk about the Midwest, where everyone thinks the monarchs have been hit the hardest. This is the region where the majority of monarchs come from, and, this is where the roundup-ready crops have led to declines in the amount of milkweed in farm fields over the last two decades (it has, we are not disputing this). Here is a breakdown of the trends from the map above, divided by region.
As you can see, there has been no decline of breeding monarchs in the American Midwest, and our analysis goes back to the early 1990s, before roundup crops were invented. In fact, there is no real decline obvious in any region. That unusual increase in the northwest may be an aberration, due to limited sampling in that region, but again, it is not a decline either. But the real take-away here is the lack of trends in the core breeding range of the Midwest. This is highly-relevant.
That's not all. We also had data on the amount of glyphosate (roundup) used in each county surrounding each NABA site. We then looked to see if sites where more roundup was applied had experienced greater declines in monarchs. They didn't. There was no correlation between roundup use and monarch trends.
Once again, no one is disputing that the advent of roundup-ready crops has led to a decline in milkweeds in farm fields in the Midwest. We know this has happened over the last 20 years, and there is solid data on this. However, the only logical explanation for our results then, is that the monarchs in the Midwest are apparently able to do just fine without that milkweed. In other words, they were never as dependent on it as we all thought. Yes, we ALL thought this. Even a lot of scientists I know had come to believe that the loss of farm-field milkweed was tragic, and there have been oodles of studies evaluating and touting this loss. And, this is the main reason why some scientists believe we must replace that lost milkweed. But in the end, it looks like the roundup-ready crops were really a red herring.
Finally, here is another astounding bit of evidence we uncovered and presented in the paper, just to cement this case. Since we had access to the NABA surveys, we (again, Mike really) examined trends for all of the other butterfly species that have been tracked over the years (>450!) to compare how the monarch is faring against everyone else. Below is a chart that shows the overall monarch trend (using all NABA sites combined) compared to every other butterfly species. We highlighted 3 well-known species for visual comparison.
This is called a histogram chart, and it shows the trend for each butterfly species combined; the higher the bar, the more species in that bar. All species to the right of zero are increasing in abundance, and all species to the left are decreasing. The monarch is pretty much overlapping zero, or even slightly above it. So really, this means that there are a lot of species that are faring much worse than monarchs - even the painted lady!
OK, so what does all of this mean? The collective evidence we amassed in this paper all point to one conclusion - that despite the declines happening at the Mexico winter colonies, the monarchs appear to be faring quite well during the summer in North America. This means that they are rebounding each year from these recent winter lows. Or in other words, the declines in the winter are not sufficient to affect the population throughout the rest of the year. To be clear, if the winter colonies eventually disappear, then that's a different story. But we can at least say that right now, the breeding population is not affected by the events happening in Mexico.
So now let's talk conservation of monarchs, and here's where it can get uncomfortable. And to be clear, I am all for protection of nature spaces and insect habitat! Now I realize that there are are currently thousands of people who have dedicated themselves to saving the monarch, and there has literally been millions of dollars spent on monarch conservation efforts. And as we speak, the governments of Canada, and the U.S. are deeply involved in "recovery plans" for the monarch in both countries. From the get-go, all of this was based on the assumption that the declines in Mexico were absolutely catastrophic to the population. Everyone (scientists too) also believed that the loss of agricultural milkweed meant that the only way to save monarchs is to plant more milkweed. But yet, our results show 1) the declines in winter do not reflect what is happening in the summer, and therefore are NOT catastrophic, and 2) by extension, there IS enough milkweed and habitat now to maintain a robust summer population.
I know, you've been hearing for a long time that more and more monarch habitat is disappearing every year. It is certainly true that traditional butterfly-friendly landscapes, like prairies, are disappearing (which is very tragic), but it is also true that we humans are also creating habitat for monarchs too, just by our own modifications of the landscape. For example, every time a forest is cut down, this inherently creates an open landscape, and a flush of new wildflower growth, which is perfect for butterflies. And if you think about it, powerlines are great places for milkweed, which thrives in disturbed land. Or how about cow pastures? These are also great for monarchs and milkweed. And the list goes on. So what I'm saying is, that the monarch is an adaptable critter, and it is probably able to utilize all of the habitats that we humans are already creating and modifying for our own purposes. Or in other words, human habitat IS monarch habitat.
OK, there is a lot to digest from all of this. I encourage you to download the paper and read it carefully to get the full story. Here is the link again. I really just covered the main points here. Let me finish by leaving you with some tough questions to ponder going forward. And some readers might be thinking these already, after reading and digesting all of this new evidence. The first is a biggie and is kind of the overarching question. This one may keep you up tonight.
- So do monarchs need to be saved at all?
- Are ANY of the current conservation efforts helping?
- Should we be spending millions of dollars on conservation of monarchs right now?
- Were the governments of Canada and the U.S. justified in declaring the monarch in need of federal protection?
I'll leave you to ponder and discuss. That's all for now.
Direct link to this blog entry: