Bombshell new study on monarch declines is published (sort of, maybe)
Last month there was a new study that came out (online) in a crazy new scientific journal, and that "appears" to upend everything we thought we knew about monarchs and their population. However, whether or not this study is valid is another question, which I'll get into in this post.
The paper was titled, "Monarch butterfly and milkweed declines substantially predate the use of genetically modified crops", and was authored by John H Boyle, Harmony J Dalgleish, and Joshua Puzey, who all appear to be from the College of William and Mary in Virginia. Before I get into the meat of this study, let me first discuss where it was "published", because that is just as interesting as the paper itself. This study was "published" in the new scientific journal called BioArchive - link here to the paper. The reason I put quotations around published is because it sort of is, and it sort of isn't. In other words, this may or may not be considered peer-reviewed science. Let me explain.
Normally when a scientist wants to publish their findings in a scientific journal, they submit a manuscript to the journal editor, who then sends the document to at least 2 outside reviewers who evaluate it. They then submit their recommendation to the editor, and based on these reviews, the editor decides whether or not it is acceptable for publication, and if so, which things need to be changed, etc. Typically, there are lots of things that the reviewers think need to be changed. If both reviewers (or sometimes just one) think the paper is wrong, then the paper gets rejected, and therefore not published. This is the way "peer-review" works in the sciences, and it has always been this way - except for now. There are some new "journals" now like this one that allow scientists to upload their draft manuscript directly to the journal, where it gets immediately made public - before any peer review. These journals call themselves preprint servers. The idea (I think) is that they allow for "greater transparency" in the peer review process, meaning that they are supposed to provide everyone the ability to make public comments on the document. In fact, on the page with this paper, there is a comment section at the bottom. This sounds like a good idea on paper, but what's more important is that the manuscript has never been formally reviewed. What if the data were analyzed incorrectly, and the conclusions were just wrong? This is why we have peer review - to prevent bad science from getting to the public.
Here is the description of the journal from its own homepage:
"bioRxiv (pronounced "bio-archive") is a free online archive and distribution service for unpublished preprints in the life sciences. It is operated by Cold Spring Harbor Laboratory, a not-for-profit research and educational institution. By posting preprints on bioRxiv, authors are able to make their findings immediately available to the scientific community and receive feedback on draft manuscripts before they are submitted to journals."
And here is their catchy youtube promo...
Don't get me wrong, I'm not saying here that this paper is bad science, in fact I think there is some valid stuff in it. But since it never went through any formal peer review, it remains unclear whether this study would have even seen the light of day (i.e. it may have gotten rejected), especially because it appears to go against a lot of prevailing theories.
The bottom line here is that this study is not peer-reviewed in a traditional sense and is therefore not (yet) what we would consider "science". It is "published" online in a preprint journal. As I understand it, these paper can later go through a typical peer review process, or the authors can elect to keep them in this stage. It's weird.
OK, now let's get to this bombshell of a study.
The goal of the study appeared to be to try to identify long-term patterns in monarch and milkweed abundance in North America, and by long-term, I mean really, really long-term - like in the last century! The authors appeared to be keenly aware of the ongoing efforts to sort out the population status of monarchs, and they also were aware of the issues and controversies regarding roundup-ready crops. They seemed to have conducted this study as sort of a test of the idea that the introduction of GMO crops in North America in the 1990s is responsible for the decline in overwintering monarchs. I'll admit this is a pretty ambitious goal.
Now, where did they get their data, especially for such a long time frame? Since citizen science wasn't around back in the early 1900s, these authors turned to a very radical (and probably underutilized) source of data - museum records. Museums these days have electronic databases of all of their insect specimens, even those collected decades ago. And, these are all downloadable nowadays, so the authors simply downloaded all records of all butterflies and moths (Lepidoptera) from 1900-2016, from something called the Symbiota Collections of Arthropods Network. This amounted to 300,000 specimen records in North America, and of these, there were 1191 records of monarchs. Keep in mind these are records of an actual specimen somewhere in a drawer in a museum, so someone had to have captured, pinned and donated the specimen to a museum for the record to exist. The authors then wanted to know how many monarchs were collected on any given year, and if this pattern has changed. However, they recognized that the amount of "effort" spent collecting probably varied from year to year, and they were right. Below is a graph showing the number of specimens (all butterflies) collected per year.
So to determine how many monarchs were collected per year, while at the same time taking into account this changing collecting effort, the authors calculated the yearly proportion that were monarchs, out of the yearly total collected. Interesting approach.
They also did the same thing for milkweeds. They apparently downloaded a similar set of records for all plants (which was much larger - 5 million records), and then calculated the annual proportion of milkweed plants, based on the total for each year.
So in the end, the authors had an estimate of monarch abundance per year, for the entire eastern range (they only focused on the eastern population here), going back to 1900 - very clever. I should mention too that they also had information on the location of each monarch specimen, and milkweed record, and with this they apparently did some interesting analyses of landcover associated with each at one point in the study. I'll get to that later.
So, for the initial goal of looking at the long-term trends in monarch abundance, I'm going to paste below their main figure. This shows the combined data for milkweed and monarch abundance since 1900. They use the term relative occurrence as the index of abundance. Take a few minutes to soak this in, because there is a lot going on in this graph...
Let's start with the monarchs. This graph is pretty messy, with points all over the place, which makes it hard to see any real pattern. However, the authors used a graphical approach called "smoothing" to create a line through the points; this line tries to find the middle of the points and it tries to show how the middle changes over time. This is a very simplified description, but the point is that the line through the points (see what I did there?) attempts to show the pattern over time.
From this graph the authors appear to conclude that the monarchs generally increased in abundance from 1900 to 1955, and thereafter they began slowly declining. And, they point out that this decline was apparent well before the introduction of herbicide tolerant crops. The blue line represents the year in which 50% of all crops were GMOs. Their reasoning in this appears sound. However, you also have to admit there is a heck of a lot of scatter to these points, which means the data is inherently messy. Also, keep in mind that the number of monarch specimens was 1191, and these are spread out over 116 years (so about 10 monarchs per year). On the other hand, you also have to admit that it is very cool to be able to look at trends in monarch abundance over the last century!
Throughout this paper, and even in the title of the paper, the authors appear to make the case that these data show that monarchs were declining well before the introduction of GMO crops, and this is the bombshell revelation that I talked about. In other words, GMO crops has nothing to do with the recent declines of monarchs (at the overwintering sites). Since their milkweed data (the green dots) also shows a remarkably similar trend over time, they argue that the monarch abundance in eastern North America is simply tracking the changing abundance of milkweeds. In fact, their estimate of milkweed abundance was correlated with the estimate of monarch abundance. And they suggest that the milkweed abundance seems to be changing on the landscape probably because of human changes to the landscape. For example, they suggest that the increase in milkweed (and monarchs) early in the century reflects the conversion of eastern forests to farmland. They did not have many ideas about what is causing the gradual decline in milkweed abundance since the 1940s, only that the makeup of farms has changed - the actual number of farms has declined over this time, because the farms are bigger. They suggest that somehow this is changing the landscape for milkweeds.
My thinking on this graph, which was the heart of this paper, is that it is certainly messy, but it's hard to argue with their milkweed data, which shows a pretty clear and tight pattern; this would be hard to come by randomly (i.e. it's tracking a real biological phenomenon). Plus the milkweed estimates are based on a tonne of records - way more than the monarchs. So I think I tend to belie the milkweed data and the pattern it shows. I'm less certain about the monarch data, but I'll admit it is fun to see data - any data - on monarch abundance going back this far. I should point out that the authors appeared to do some comparisons of their monarch abundance data (from the last 20 years) with some other known datasets, as evidence that these data were sound. It looks like their data was correlated with the estimates of egg and larval densities in the east, based on the MLMP citizen science program, and their data was almost significantly correlated with the number of adult monarchs counted by the NABA program. So to be fair, it looks like their data on monarch abundance is not that far off. Which means, that we should probably consider the rest of the story here to be valid (i.e. GMO crops not being responsible, etc...).
The other thought I had was that throughout this paper the authors put a lot of effort into highlighting the "decline" of monarchs over that last 50 or so years, which seemed odd to me. It may not have been fair for the authors to focus solely on the monarch decline in the last 50 years, because this came after there was an increase in monarchs! If you look at the numbers of monarchs on this graph for the last 20 years, and compare it to the numbers for the first 20 years, they are pretty much the same. I'll show you a graph below that I made to prove this. I took their graph and made a copy of it, which I used to recreate their original data (you can do this with special software). My copy of the data is not an exact match, but it's pretty close. Then, with these data I calculated the average number of monarchs seen in the last 20 years and compared it to the first 20 years, graphed below.
From this graph you can reasonably conclude that the number of monarchs we have today is pretty much the same as it was in the early 1900s. Don't worry about the values on the y axis, the point is that the two bars are similar (and statistically undifferent). This is a very important point, and one in which the authors did not make. In fact, they appeared to focus only on the decline, probably because this makes the story sound more dramatic (no one wants to hear that the monarchs are doing fine!).
As I said earlier, it's probably not fair (is that the right word? or maybe "appropriate") to focus on the decline, which came after a substantial increase, especially when you wind up with the same number as you had when you started, as my graph indicates. An analogy to this would be if you go to Vegas to gamble, and you start with $100 in your pocket. Say you win big for a while until you have about $1000 in your pocket, but then you loose big on the last day, and you wind up going home with $100. Then your spouse asks, how did you make out? You could say, I lost $900, which would be true. Or you could say I broke even, which is also true.
There were other parts to the study that I don't have the time to get into, including some interesting bits about glyphosate use (roundup) and whether or not this predicted monarch abundance. Perhaps you should read the paper to learn what they found here...
I think at this point I've probably covered the main talking points in this bombshell of a study. This paper, if it ever does get officially published, will surely upend a lot of theories about the current status of the eastern monarch population. If it remains as is (on a preprint server), then I'm not sure this counts as science. If the authors are reading this - get this thing published!
Til next time...
Direct link to this blog entry: