Recoveries of tagged monarchs in Mexico - what do they really mean?
Summer is just about here and by now everyone has seen lots of adult monarchs in their yards and maybe eggs or caterpillars. Before you know it, thoughts will start leading toward the fall migration, and starting our tagging efforts! Well, before that happens, I'd like to spend today talking about the data that is generated by tagging, and tell you about some recent research focused on the tagging and recovery data. So this is a post that needs to be read by anyone who has ever tagged a monarch. Most taggers generally know that these tags are "used to help track the migration", but really, this is an oversimplification. In reality, it is much more complicated than that, as you'll see from this post.
First, let's start with the basics, just in case any readers here have never done this. Tagging monarchs essentially involves placing a numbered sticker on the wing of an adult monarch, as pictured above (from MonarchWatch.org), and then releasing it. If that monarch is spotted or found again by someone who can read the number, they then report where the monarch was found. This allows us to place two dots on a map (release location and recovery location), and then draw a line connecting them. This is a rather archaic technology, considering the recent advances in GPS tracking, but on the other hand, this is the way that the Mexican overwintering locations were tracked down years ago - by drawing lines on a map. So it is really thanks to a lot of tagging efforts by citizens back in the 60s and 70s (who were recruited by Fred Urquhart), that we now know where that all-important location is.
I bet a lot of folks reading this have even had one of "their" monarchs recovered in Mexico. If you tag long enough, this will happen eventually. My friend Don Davis is a legend for having lots of tag recoveries in Mexico! Anyway, this is usually something that is celebrated, and you even get a certificate in the mail when this happens. People like to post their recoveries on their favorite monarch facebook site too. I've always been a little uneasy about this celebratory nature of it though, since it seems to turn this very important scientific endeavor into something akin to a lottery.
After Fred Urquhart stopped running it, the monarch tagging (in the eastern N. American population) has been run by MonarchWatch for the last 20 years or so. I've always said that these tagging data are very important for the conservation efforts, since they tell us where the migration flyways are, and where in these flyways are the more critical locations. In fact, I even published a paper stating this. But recently, I've begun to question whether these data are as useful as I thought. I explain below.
Last year, a group of scientists, led by folks from MonarchWatch, conducted a formal analysis of 18 years of tagging data, the results of which were published in the journal, Frontiers in Ecology and Evolution. This paper is fully online, and a link is here. The intent of the paper appeared to be a test of something called the "migration mortality hypothesis", which is an idea that basically came from my lab. Essentially, it goes like this: since the size of the Mexican overwintering colonies has been declining over the past two decades, but, the counts of adult monarchs in the breeding range has not declined, therefore, there must be increasing mortality during the fall migration that is leading to the Mexico decline. Incidentally, this hypothesis has a fair amount of support from a variety of different studies now. Anyway, the MonarchWatch paper was intended to be a kind of test of this hypothesis. The researchers examined the rates of Mexico recoveries (from the tagging records) over 18 years. The rate is simply the number of tagged monarchs that were recovered in Mexico, divided by the total number of tagged monarchs. Typically it is something like 1-2 percent. The researchers figured that if migratory mortality were truly increasing over time, then there would be a corresponding decline in Mexico recovery rates over time. They did not find this, based on their own analyses of the tagging data. Their analysis showed that recovery rates have varied a little, but overall showed no real trend. Therefore, they concluded that the migratory mortality hypothesis was busted (or "debunked" as I read in one press release).
Now, here's where it gets interesting. After they published their paper, another group of researchers - James Fordyce (U. of Tennessee), Chris Nice (Texas State), and Matt Forister (U. of Nevada) - took issue with their study, and then published something called a "rebuttal" paper. This is kind of like a small study that refutes the first one, or maybe provides a different look at the data. These typically get published alongside the first paper, so that people can read the arguments from the rebuttal authors, and then make up their own mind about the original paper. Sometimes, the rebuttal paper is then followed by a counter-rebuttal by the original authors. I happen to know that there will be no counter-rebuttal forthcoming for this paper though. Anyway, this rebuttal of the MonarchWatch paper is now fully online, and is linked on the same page as the original paper - link here.
Let me give you the gist of this rebuttal. The Fordyce et al. group took issue mainly with the way that the MonarchWatch group statistically analyzed the recovery data. There were a number of different tests performed in the MonarchWatch paper, but the most important was the test that examined recovery rate over time. The original group had basically run a correlation test, to see if the recovery rate per year has changed over time (comparing recovery rate versus year). This seems fair at first, but the Fordyce crew said that this approach does not take into account the vast differences in population size each year, which may affect how many monarch monarchs are tagged, and/or recovered. So, they ran their own analysis using the same data as MonarchWatch (which had been posted in the original paper as tables), and their analysis took into account this size variation. In essence, they still were examining if the recovery rate has changed over time, but now this approach factored in the population size variation.
Shockingly, their analysis showed that tag recovery rates in Mexico have steadily increased over the last 18 years! What? Yes, it's true. I recently reached out to my friend Matt Forister, who pointed out to me what they did. I just crunched the numbers myself and it seems solid.
To be honest, I'm just as shocked as you are about this. On the one hand, this new information seems to refute the original paper's analyses (big time), so in other words, they were wrong. On the other hand, this this new evidence certainly seems to go against the migration mortality hypothesis too. How can recoveries in Mexico be increasing - can the monarchs' fall migration be getting easier over time? No, that's just stupid.
In their rebuttal paper, these authors made a few suggestions for why they think this might be happening. For one, they pointed out that the recovery data from Mexico is dependent on people looking for tags (more on this below), and, we really don't know if this "search effort" has increased over the years. If it has, then this might explain the apparent increase in recoveries. But, Taylor and colleagues have argued very firmly that they don't believe the tag search effort has increased, so there's that.
They also pointed out another possibility, and this is one that I am inclined to believe. They suggested that there might be a general increase in mortality of monarchs at the overwintering colonies (due to stress from in increasingly harder migration), which is leading to more tags being found over time. This makes a lot of sense to me, but first, let me unpack this idea a bit further. Something that most people seem to forget, especially when they get one of these celebratory certificates, is that a tag recovery in Mexico means that your tagged monarch died there! And, therefore it did not remigrate north in the spring, have offspring, or contribute to the overall population at all. Tags are found by locals who search through the forest floor, surveying the dead monarchs, not by looking through the living clusters with binoculars. In fact, the local monarch guides are the ones who usually find the tags. There was a neat essay a number of years ago on this, which was posted on Journey North (link here). Reading it, you can see that the locals essentially sell the tags back to MonarchWatch. So, this means there is a monetary incentive for them to collect these tags. Below is a neat picture from that essay, showing two youngsters holding a bunch of tags, waiting to get paid!
So, when you consider that tag recoveries are essentially deaths in Mexico, you can start to picture what might be happening - increasing tag recoveries over time really means increasing winter mortality in Mexico. Probably not a good thing.
Now, here's a question that will really throw you for a loop - are the tags themselves leading to the monarch deaths? Think about it. We already know that the stressors monarchs face during migration are immense - like cars, loss of nectar, storms, etc. What if the weight of the tags, even if it is incredibly minor, is causing monarchs to burn slightly more energy during flight every day, so that they arrive in Mexico with not quite enough fat to survive the winter. Or, maybe the minor weight of the tag causes the monarchs to not flap as efficiently as it would have normally. Or, maybe the tags are leading to slightly, but chronically-elevated metabolic rates. Or, for all we know, there is a behavioral issue at play here - maybe the white tags cause the tagged monarchs to be shunned by their untagged friends, and the tagged monarchs are then prevented from joining the rest of the monarchs in the safe tree clusters. Who knows? The point is, we really don't have any data on the effect of tags to monarch physiology, flight mechanics, or behavior to say anything about this, but, given this new evidence (about increasing mortality at the winter sites), I might want to take this issue up in my own lab, and maybe other scientists should too.
OK, so to sum up this post, we now have two competing groups of researchers, who have each looked at the rates of monarch tag recoveries over time. A simplistic analysis of the data from the first group indicates that tag recoveries have not changed over time. A more comprehensive analysis from the second group shows that it has, which raises lots of questions, like why monarchs would be dying more frequently when they get to Mexico.
If there is one thing that this whole experience (the original paper, plus the rebuttal) has demonstrated, it's that interpreting trends from monarch tagging data is surprisingly hard to do. So the next time someone tells you that these tags "help scientists track the migration", tell them that it's more complicated than that!
That's all for now.
Direct link to this blog entry: