top of page
  • Andy Davis

Another monarch study published showing the spring migration holds the key to everything

Hello everyone,

It's been a fantastic year in the world of monarch research! In the past 11 months there have been some amazing research projects published, including two themed collections of papers - one devoted to monarch migration (in my journal - I blogged about it), and another that is being published in stages, in a journal called Frontiers in Ecology and Evolution. Currently, there are about 28 new studies published in just this collection alone! And, they're all completely online. In terms of the science around monarchs, there is just too much to talk about these days, and it's almost a little overwhelming, even for a monarch scientist.

Today I'm going to tell you about one recent paper in the Frontiers collection that really caught my eye. It was authored by a friend of mine and fellow Canadian, Tara Crewe, and her colleagues. The title of the paper pretty much speaks for itself, and gives you the gist of the paper too - Size of the Canadian Breeding Population of Monarch Butterflies Is Driven by Factors Acting During Spring Migration and Recolonization. Here is a link.

As the title suggests, this paper is all about monarchs in Canada, or the province of Ontario, to be specific. Recall from another Canadian study that came out earlier this year (Flockhart et al, where the breeding range was mapped for Canada), that southern Ontario is pretty much the heart of the Canadian monarch world. So this region is pretty important, since it holds most of "Canada's" monarchs each summer. And, we also know that this region contributes a lot to the overwintering cohort in Mexico - the legendary tagger Don Davis lives in southern Ontario and he has a ridiculous tag recovery rate.

The question Tara and company were trying to address here, was what is it that causes the Ontario breeding population to be high in some years but low in others? Is it the weather? And she also wanted to know if the high years in Canada then translate to large overwintering colonies in Mexico? And finally, she wanted to know if large Mexican colonies lead to large Canadian breeding populations?

To answer these questions, the authors used a very unique dataset, which I don't believe I've ever seen used so far in the monarch research world - the Ontario Butterfly Atlas. A lot of provinces and states have these, either for birds, butterflies, and I've even seen some for herps! These are citizen science programs where volunteers go out and census specific coordinates in the region, and record all species (of butterflies, in this case) they see. The information gets compiled by someone and then once all coordinates have been surveyed, they can make some groovy maps showing the range of each species. Some atlases are kind of a one-time thing, like once the coordinates get checked, then that's it. Others are more of an annual thing, where as many sites are checked each year as possible. It looks like this is what the Ontario atlas is. Incidentally, I just googled this and found a link to their homepage here. So it looks like this atlas program has been running for a number of years now, which is great because that allowed Tara to ask these questions about the annual changes in numbers.

There was a map of the sites provided in the paper, but buried in the Supplemental material. Here it is:

I should be more specific - actually, this map shows the locations where monarchs were recorded (orange dots) as well as places where monarchs were not recorded (grey dots). Remember the surveyors record all butterfly species seen in their surveys. It looks like on average, each survey recorded about 6-8 different butterfly species. Judging by this map, I'd say they had a lot of data to work with for this paper!

So for their analyses, the authors calculated an annual index of monarch abundance for all of Ontario, using these data. There was a detailed statistical model used for this, but for simplicity, we can just call it an average. And also, they did this for all of the most common species recorded as well. Once that was done, they plotted these indices over time to see how they have changed (or not). Below is a series of charts showing these indices, for all butterfly species, including monarchs (Danaus plexippus).

The y axes of these charts are a little hard to explain - they are not really the average number of butterflies, but more like an index that the computer model spits out based on all of the information it has for each year. So don't hurt your brain trying to figure these out. Just know that the higher the number (even if it's negative), the more butterflies that year. Also, you'll see that there are some "error bars" on top of each annual point - these show how much statistical variation there was each year - sort of like, how confident the computer models were about those averages. Notice that some species have small error bars compared to monarchs, which have large error bars. That means monarch abundance tends to be highly variable each year across the various survey sites.

One key thing to point out with the monarch graph, is that it's not showing a decline in abundance over the 15 years. More on this later.

The next step was to ask what is driving the year-to-year changes in abundance of monarchs in this dataset. To do this, the authors compared these numbers with a number of other datasets over the same time frame. First, they looked to see if the size of the Mexican overwintering colonies predicts how many Ontario monarchs there are - essentially, they looked to see if there is a correlation between the two datasets. There wasn't, or at least, not a strong one. Below is a graph comparing the two.

In the graph above, the number of monarchs counted in Ontario is the orange line (the same trend shown in the prior graph). Above that in blue are two lines representing the overwintering colony size over time. There are two lines here because one (the long dash line) is a computer model prediction. The other (short dashes) is the actual colony size estimates. Again, don't worry about the y axis too much. The important thing to point out here is that the two datasets don't really match up. This means that the size of the overwintering colony in any given year has no bearing on how many monarchs eventually arrive in Canada to breed, and/or how big the Canadian breeding population is. This is one conclusion of the paper.

Next, the team wanted to know if the Ontario counts could be predicted by something else, like the weather during the spring migration. They had an idea this would be the case, because of some previous studies, which I'll come back to. So, they compared the Ontario data with data on the number of days in the spring (April and May) when the temperature was optimal for monarchs (21C).

This graph also has some wonky axes, even for the x axis, so don't worry about the actual numbers. The key thing to see is that there is a (very) positive correlation between the annual temperature in April and the size of the Ontario breeding population. That's right, in April. Like when the monarchs are way south of Ontario in Texas, and are in the early stages of their spring recolonization phase. The authors here didn't say much about this, but it stands to reason that the higher temperatures in April would be conducive to either faster milkweed growth, or would help monarch movement - monarchs can only fly if the temperature is over 13C or so. Or, the warmer temps might speed up larval development. Or maybe it's a combination of everything. Whatever the reason, it looks like this right here is the key to everything with the Canadian monarch population - the success of the early spring migration. Everything hinges on this. Incidentally, the authors also looked to see if the weather in the breeding period is correlated with the number of monarchs - it wasn't.

Another interesting set of analyses done here was to compare the annual variation in monarch numbers in Ontario with the annual size of the other butterfly species shown in the above charts. They specifically wanted to see if there was more correspondence between monarchs and the sedentary species (the green graphs) in any given year, or with the migratory species (the orange graphs). If there was stronger correlations with sedentary species, that would mean the summer conditions are what's driving the monarch numbers. If there was a stronger relationship with migratory species, that would mean that factors outside of the summer are what's important for monarchs. This is a little bit convoluted, but this was still an interesting way to ask these questions. In short, it looked like there was a stronger relationship between monarchs and the other migrant species. So, this makes sense considering the earlier conclusions about the importance of the spring conditions.

The spring conditions seem to be all-important for monarchs. Actually, this wasn't the first time this very thing has been found. Two other studies have also concluded this same thing, using other datasets. They both came from Elise Zipkin's lab at Michigan State University. Her lab does some very cool statistical analyses of monarch population data (and other topics). And, I just saw that her lab was awarded a big grant to continue studying monarch trends (link), so we should soon see more from this group. Anyway, in two other studies, her group showed that, statistically speaking, the best predictor of how many monarchs you see during the summer in the Midwest is the success of the spring recolonization early in April and May. I blogged about these last year. And, that success highly depends on the weather during the spring. So now we have three different studies, using three different datasets, that all point to the same thing - everything hinges on the success of the monarch spring migration, and, that success hinges on the spring weather.

OK, I know what you're thinking - so if the spring weather is so important for producing large summer populations (in either the US or Canada), then what can I do to help the monarchs? Is my garden helping? Well, unless you can change the weather, then the short answer is, there isn't much you can do. Sure, you can plant some milkweed if it makes you feel good - after all, it sure is fun to see the monarch caterpillars on it. And to be fair, monarchs do need milkweed in the spring to reproduce. But keep in mind that the majority of research, including this current paper, shows that the summer population of monarchs, both in the US and in Canada, is pretty stable - meaning the monarchs are not in trouble (despite what you've heard), and, they are not dying from a lack of milkweed. Think about this - this very paper showed that the average number of monarchs seen in the Ontario Butterfly Atlas has not gone up or down in the last 15 years. This means that the monarchs have been able to successfully recolonize their Canadian population each year, using the current supply of milkweed.

Finally, I'll pass on some thoughts directly from the lead author here. I corresponded with Tara after the paper came out, to get her thoughts on the results. She told me, "I think the main point to take away is that spring migration/recolonization appears to be the bottleneck - without good conditions, the population won't have the potential to recover from lows."

Excellent study - well done, Tara and company.

That's all for now. Thanks for reading.


Direct link to this blog entry:



The science of monarch butterflies

A blog about monarchs, written by a monarch scientist, for people who love monarchs

bottom of page