top of page
  • Andy Davis

An objective look at the 2016 breeding season, using MLMP data

Greetings all,

Since the monarch breeding season is starting to wind down now I thought I'd spend some time in this blog entry trying to make sense of the breeding numbers for this year. So for this blog get ready to see some cool graphs and data!

Let me first set the stage for you - in between chasing the kids this summer I was reading all of the online reports and emails about the numbers this year - the number of eggs and larvae, that is. It seemed like there were some areas where people saw lots and some where there weren't very many. From reading these reports it was not really clear (to me, anyway) how the breeding season stacked up this year. Perhaps other people felt the same. Today I'm going to hopefully set the record straight, and do so in an objective, scientific fashion.

Here's where the science comes in - for this blog I did some digging into one of the best sources of data on summer reproduction, the Monarch Larva Monitoring Project ( As most long-time monarch folks know, this program, run by the Oberhauser lab at the U. of Minnesota, asks volunteers to monitor patches of milkweed on a weekly basis over the summer, and record the number of eggs and larvae they see. The data then get posted online and compiled to make nifty seasonal graphs of monarch abundance.

Importantly, the data from all sites over all years of monitoring is archived at the mlmp website, so that users (like me) can go in and query the database to pull up graphs from a given state. The program can compile all monitoring data from each state into one graph. Or I think you can also go in and pull up data from any one monitoring site within the state as well. Either way, it's a very slick website, and Karen's people keep making it better each year. Kudos to them.

So I asked the MLMP website to give me all graphs for 5 states that I chose to best reflect what's going on with the summer reproduction - Minnesota, Michigan, Wisconsin, Pennsylvania and Iowa. The graphs it spits out look like the one below, which is from Minnesota in 2016 (you have to generate the graphs one at a time).

So these graphs tell you how many eggs and larvae were observed on plants over the summer, for the entire state. In this particular state, the graph says there were 35 site monitored. Now as I understand it, the important thing to look for in these graphs is the peak density of the final generation, which will be the cohort that ultimately heads to Mexico. In this graph, it is the bar dated 8/14, which reads about 0.08 monarchs/milkweed. This is the way that other monarch scientists who have worked with these data have handled the numbers - by focusing in on that one bar at the end of the summer.

I went through each graph for each of the 5 states I mentioned (I also picked these states because they had the most data), and I wrote down the peak number for each year, based on my eyeballing the graph. Forgive me for not getting my ruler out, but I think I did pretty well. Then I started crunching some of the numbers and making my own graphs from these numbers to try to figure out my original question - how did 2016 measure up?

I'm going to paste here the 5 graphs I made based on the MLMP data. In each of these graphs, I plotted the peak monarch density per year, for as many years as there were data within each state. These graphs will essentially tell us how the abundance of eggs and larvae have (or have not) changed over time within each state, and also allow us to compare this year with all prior years.

OK, I know what you're thinking - what's going on in Pennsylvania this year? Was that last point a mistake? I thought that too, so I went and looked at the raw numbers from that state (the mlmp website allows you to download the raw numbers), and sure enough, in one of the final weeks of monitoring (Sept. 11), there were 235 eggs and larvae (150 fifth instars!) seen on 90 or so plants. Given that these were mostly fifth instars, they would surely be part of the migratory cohort, so this number seems right.

Speaking of the Pennsylvania numbers, it's interesting to note that in 2015, all other states saw pretty good monarch numbers, yet in Pennsylvania, there were low numbers seen. Weird.

Next I'm going to show you a graph where I tried to cram all of the data from these 5 states into one plot. Here, I calculated the average monarch larval density per year, from the 5 states - so each point below is an average of 5 numbers.

Cool, eh? So this graph seems to indicate that the monarchs did pretty well this year, compared to previous years, at least for the 5 states I examined. Incidentally, I also calculated the average without the Pennsylvania data, in case someone out there complains that those Pennsylvania numbers from 2016 are throwing things off with their apparent caterpillar bonanza (how dare those Pennsylvanians find good numbers of monarchs, anyway). Someone could also complain that this state is not in the 'CORE' breeding range of the Midwest. To make everyone happy, here is a graph without those pesky PA numbers...

Now, getting back to the original question - how did the 2016 breeding season measure up? Well, it looks like the answer to that question depends on what dataset you look at. If you just look at larval data from the midwestern states I chose (MN, MI, WI, IA), the peak numbers this summer were down from last year, and lower-than-average for the region (the overall average for the 4 states in the graph above is 0.29). But, if you include other states not in the midwest (Pennsylvania), the overall numbers this year were better-than-average.

There is another thing to note about the last graph above, which is the long-term trend over the past 20 years (for only those midwestern states). From this graph it looks like the numbers of monarch eggs and larvae in the past 5 years (2012-2016) is not that much different than those from the early years (1997-2001). When I crunch the numbers I find the average density for the first 5 years was 0.23 monarchs/milkweed, and in the last 5 years it was 0.18 monarchs/milkweed. OK, so there is a small change. But if you do the same thing and include the Pennsylvania data, the first 5-year average is 0.24, and the last 5 yr average is 0.31. So the numbers have increased.

Interesting stuff, indeed. I'm going to leave it here, and maybe come back to these data in a future post. It's been really fun to play with these numbers. I think we should all thank the folks at the Monarch Larva Monitoring Project for coordinating these data, and for making such a slick website. We should also be especially grateful to those hard-working citizen-scientists who collected the data!


Direct link to this blog article:



The science of monarch butterflies

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

bottom of page