Beesource Beekeeping Forums banner

1 - 5 of 5 Posts

·
Registered
Joined
·
6,039 Posts
Discussion Starter · #1 ·
I've been playing around with this Excel model he has and it's really very interesting and thought provoking.

My question is in treatment efficacy. Mainly, if I'm dropping in Apivar, would you call that a one time, say, 90% kill or since it's in there for six weeks is it a 90% kill in three of the two week boxes? Thoughts? It doesn't kill 90% the day it goes in, but it also could conceivably help keep the effective drift down over that six week period.

I'm not really using the data to make any decisions, but it is changing my thinking a little bit when starting to consider how big drift might be when it comes to re-infestations later in the season.

Thanks!
 

·
Registered
Joined
·
2,032 Posts
I've been playing around with this Excel model he has and it's really very interesting and thought provoking.

My question is in treatment efficacy. Mainly, if I'm dropping in Apivar, would you call that a one time, say, 90% kill or since it's in there for six weeks is it a 90% kill in three of the two week boxes? Thoughts? It doesn't kill 90% the day it goes in, but it also could conceivably help keep the effective drift down over that six week period.

I'm not really using the data to make any decisions, but it is changing my thinking a little bit when starting to consider how big drift might be when it comes to re-infestations later in the season.

Thanks!
I'm not 100% sure I am grasping your question, so ignore me if I am off the mark. I have never thought about this before. My first thought is that you get the 90% kill as of the END of the effective period of treatment, right? So "treatment date" input should be input as of the last date in the treatment cycle? From that end date, you start your drift/growth clock over?
 

·
Registered
Joined
·
6,039 Posts
Discussion Starter · #3 ·
I think you're right PSM, I guess what I'm getting at is how to model that. Do you drop the 90% in for when you pull it off? Or do you put in something different. I'll have to look again, maybe you could zero out the drift at that time.
It's not super important practically. But just playing around with it and it makes a huge difference how the end of the year ends up (in the model).

For instance, I usually pull honey around July 1 and put Apivar strips in then. If I put a 90% kill in for mid-August when I'd pull them. It shows a crash because during that entire 6 week treatment period the mites are still building up to "crash" later in the summer. In short, putting a 20% kill in the period before the 90% "pulling" of the strips, and there's almost no mites to speak of in the model. This particular model also has OAV cleanups later in the year. This is all theoretical, I'm just curious as to how it should be modeled or if someone happens to know how Randy would suggest using it. I wonder if I could search up a video of him talking through it somewhere.

15% treatments are me modeling an OAV every five days for five treatments in a colony with brood. The 90% is me modeling Apivar. The 80% treatments are me giving them two broodless or nearly so broodless treatments in the fall. Again, this is all just me messing around with the model.
63929



Here's the same exact model but leaving out the little 20% before the Apivar treatment.
63930


Mathematically, I understand why this happens. The period leading up to the 90% treatment is peak mite population and so as far as the model is concerned, the 3000 mites in the hive are still going to town for the six weeks before I "kill" them in August. But practically... this isn't really how it happens as hopefully, the Apivar is killing the vast majority of every mite that comes out on a bee at that time.

Again, I totally understand that this is a model and all that. I'm just kind of wondering how to most accurately model the system theoretically. :)
 

·
Registered
35
Joined
·
2,089 Posts
Again, I totally understand that this is a model and all that. I'm just kind of wondering how to most accurately model the system theoretically. :)
The problem I found when trying to model mites was we tend to think in terms of kill rates, but for the math to work, you have to instead look at survival rates.

If you are expecting a kill rate of 90% over 3 periods, then you are expecting a survival rate of 10% over those same 3 periods.

If you model the kill rate at 53.5% per period, that's a survival rate of 46.5% on each period. 0.465 ^ 3 = 0.10, so with 46.5% surviving in each cycle, after 3 cycles, only 10 percent are left.

When I wrote up models for predicting mite levels, I found that focussing on kill rates made the math unbearably complicated and it just didn't work. As soon as I switched my mindset to survival rates, suddenly the math works well and is strait forward.
 
1 - 5 of 5 Posts
Top