Hi folks,
I've been up to my ears in other projects in my lab lately, so haven't had much time for a new blog. Fortunately, there has not been any huge developments or publications involving monarchs in the last month. In case anyone is interested, here is a recent story that was done on some of my other work (sorry for the shameless plug) - https://www.nytimes.com/2019/05/02/science/parasites-beetles-wood.html
One thing that is of interest right now in the monarch research world is this story that came out last year that I blogged about at the time. Recall the controversial study that first came out in that "preprint" journal - BioArxiv, regarding monarch declines and milkweed, and which was written by JH Boyle, HJ Dalgleish, and JR Puzey. Here is a link to it. It was a study that focused on museum records of monarchs in eastern North America over the last century, and it apparently had found some interesting patterns in monarch abundance over time. Recall that the main findings of the study were that monarch abundance (based on museum records) had gradually increased in the first half of the 1900s, then gradually declined from about the 1950s. They also had data on milkweed abundance, which showed a pattern of decline from around the same time period. This information seemed to fly in the face of the idea that the declines in monarchs abundance have been caused by the adoption of herbicide-tolerant crops, in the mid-1990s. I had blogged about this study when it first came out in the preprint journal - here. And to jog your memory, I'll repaste the main figure from their study below:
And recall that the y axes in these graphs are a bit unusual. The researchers in this study wanted to know how many monarchs there were each year, based on the number of museum records of monarchs each year. But, they anticipated that in some years there would be more records simply because there were more people collecting those years - in other words, the effort per year probably varied over time. To control for this (and this is what is being argued the most), the researchers used the total number of museum records of all Lepidoptera per year as a sort of control for the amount of "collecting effort" for any given year. Thus for each year they calculated the number of monarch records divided by the total Lepidoptera records - that's why the graph above says "Relative Occurrence" on the y axis. They did the same thing for the "relative occurrence" of milkweeds - relative to the total number of herbarium records - although it doesn't look like anyone is arguing the milkweed trends.
I recall being a little skeptical of this study at the time, mostly because it wasn't really published in a real scientific journal. That journal, BioArxiv, is a place where scientists can send their unpublished, and non-peer reviewed manuscripts, where (for a fee) it will get placed directly online. The idea is, that scientists can place the paper here temporarily, before they submit to a real scientific journal. Well, apparently, these authors did that. The paper later appeared in a really, really prestigious journal, Proceedings of the National Academy of Sciences (PNAS)!
Once it appeared there, it looks like it started getting noticed a lot, and then discussed - or in other words, argued about. The controversial nature of the findings seemed to be the thing that people discussed most. That, and the way the authors calculated the monarch abundance indices.
As a testament to this, since the paper was published in PNAS, there have been two "rebuttal" papers written about it. For non-academics, these are papers that scientists write when they want to challenge the findings of a particular study. The first was written by a fellow named Tyson Wepprich, and I'll link to that rebuttal paper here. He pointed out that the original study had incorrectly calculated the annual number of monarchs from the museum records. Remember, in the original study, the authors tried to correct for the variation in butterfly sampling from year to year by dividing the annual number of monarch records by the total number of ALL LEPIDOPTERA records - i.e. all moths and all butterflies. But, Wepprich pointed out that using all Lepidoptera records, including moths, is misleading because moths are usually trapped using lights at night, while butterflies are usually caught in the day with nets. He then used the same dataset and re-calculated the monarch abundance using only the records of butterflies, and then found a completely different long-term pattern of abundance than what Boyle et al had found. Below are some graphs from his paper. These graphs show the corrected relative number of monarch butterfly records (the y axis on both), but the top one shows the calculated monarch abundance based on the total butterfly records, and the bottom graph shows the calculation based on only the Nymphalidae records - that's the group of butterflies that monarchs belong to.
Both graphs show no evidence of a decline since the 1950s as was seen in the Boyle et al study, and these in fact appear to show that monarchs have recently been increasing! From reading the Wepprich paper myself, he writes that he himself is skeptical of this pattern, because, in his words, "I do not think that this reanalysis presents the true monarch trend, since it contrasts with recent declines." In other words, these data are not to be believed, since they do not show a decline! Let me come back to this later.
Now, there was a second rebuttal paper just published a few weeks ago in the same pre-print journal, by Leslie Ries and colleagues. Here is a link to that one. This paper was rather similar to the Wepprich paper, as it was arguing that the method of calculating monarch abundance was incorrect. These authors also argued that the abundance index should not have incorporated the moth records. But in addition, they also pointed out that there were other problems. For example, the original paper did not consider the timing of the museum records. They pointed out that monarchs are most abundant during the summer months, and the original paper should only have considered records of monarchs within this time frame. They also pointed out that there are some cases where a disproportionate number of monarchs were obtained in certain years and in certain locations, so that the museum records tend not to be evenly spread.
Ries and colleagues then did their own corrections to the original data and then made their own graphs, pasted below. The y axis on this graph is the relative number of monarch records, based on all butterfly records. And note that this dataset has been corrected to remove certain records.
Similar to the Wepprich graph, this one also shows a pattern where it looks like monarchs have been increasing over the last 40 years or so. But, like Wepprich, it looks like Ries and colleagues also did not believe this, and they argued that this proves that the data are not appropriate to understand monarch (or other butterfly) abundance. It's just too sketchy, in their book.
You can probably gauge my feelings about this. In both rebuttal papers, the authors (whom I respect as scientists) made the case that these data are not to be believed, because these long-term patterns don't show what we humans think they should show. I'm a little uncomfortable with this practice. I know this happens all the time in science though. It's a real problem in fact. But let's imagine for example if these corrected graphs had shown a long-term decline in monarch abundance, based on the museum records. Would they be believable then? Why then? You can see my point here. We should not really be doing this at all - interjecting our own presumptions about what a dataset or trend should look like. As long as the data are sound, then as scientists, we should believe what they show.
To be fair, I'm not saying I believe or disbelieve these data - the original or the corrected versions - I'm just pointing out that there is an element of human nature that comes into play in science, where people see what they want to see and disbelieve what they want to disbelieve.
I guess the take-home message from this whole experience - the original paper, the backlash, the rebuttals - is that the science on any given subject is always changing and evolving - hopefully in a positive direction.
That's all for now.
****************************************************************************************************************
Direct link to this blog entry:
****************************************************************************************************************