5 ms·

Ecology is incredibly difficult because everything is horrendously under-sampled. Essentially all ecology reduces to counting species over time and somehow extr

by joshvm 6y ago

Ecology is incredibly difficult because everything is horrendously under-sampled. Essentially all ecology reduces to counting species over time and somehow extrapolating to huge areas. So-called total count surveys are extremely expensive. I can only speak from a megafauna perspective (eg not insects), but there were recently some results that suggested that animal decline in Africa was overestimated. Caused quite a stir. And it makes sense when you learn how people count animals in aerial surveys - photographic interpretation is novel because most ecologists are not used to dealing with that level of data, nor the complexity in designing the camera systems, and it was impractical to do with film in the past. The gold standard is someone with a voice recorder in a plane. When you actually go through aerial photographic surveys, you spot all sorts of things that the observers missed, because they're not superhuman. That's not to say there is no decline - poaching has still decimated populations - but for some species it might not be as bad as you think. In theory the methods (things like JOLLY2) are supposed to correct for you missing a certain percentage, and distance to the animals etc, but that relies on you knowing what those uncertainties are. I don't know how insect surveys are performed, but you have to understand that you're always trying to make sweeping assumption about "the whole of the US" based on very sparse sampling. Not to mention controlling for different environments. For example in the paper you cited, they try and control for a ton of stuff like habitat. I'm not an entomologist so I won't comment on the paper beyond that, but do be aware that "standard techniques" may well be something that we've done since the 60s and nobody questions it.

DuskStar 6y ago
> but that relies on you knowing what those uncertainties are. And it also relies on those uncertainties not changing. "Elephants start spending more time in the forests and not open areas to avoid poachers" would reduce counts from airplanes without changing the number of elephants, for example.
Natsu 6y ago
For insects, there are a lot of different traps based on the nature of the insects being sought. There's a pretty good list here: https://en.wikipedia.org/wiki/Insect_trap https://en.wikipedia.org/wiki/Insect_trap As I understand the sampling, you pretty much just put the traps out and count up how many insects you get of each type. Purely anecdotally, I feel like the number of butterflies is way down from the level you'd see during, say, the 90s. It seems like I just don't see as many different types any more and even the more common types feel vastly reduced in number. I used to see monarchs and sulfers and painted ladies and red admirals and cabbage butterflies all the time, now I barely see a few cabbage butterflies or sulfers and occasionally a painted lady or two. I never get to see the less common things like buckeyes any more, which is a pity given that they're my favorite.
- 9nGQluzmnq3M 6y ago
  This is anecdotal, but during Singapore's lockdown, one of the many services cut for about two months was curb maintenance. Being deep in the tropics, all green spaces immediately turned into jungles of flowers and weeds, and the butterfly population consequently exploded: https://mothership.sg/2020/06/grass-cutting-butterflies/ https://mothership.sg/2020/06/grass-cutting-butterflies/ https://www.bbc.com/news/world-asia-52960623 https://www.bbc.com/news/world-asia-52960623 On the other hand, mosquito populations also increased and there were lots of people stuck at home as bait for them, contributing to Singapore's worst dengue outbreak in years: https://www.tnp.sg/news/singapore/three-more-dengue-deaths-singapore-outbreak-worsens https://www.tnp.sg/news/singapore/three-more-dengue-deaths-s... So at least one lesson here is that, if given an opportunity to bounce back, many insect populations can and will.
  - Natsu 6y ago
    Wow, Singapore has some really pretty ones. I've always been partial to butterflies with eyespots.
  - refurb 6y ago
    Singapore is also going through a particularly bad Dengue fever season, so there is a ton being done for mosquito control including spraying of insecticides. I have no idea how selective those insecticides are, but one could imagine an impact on other insects.
    - CydeWeys 6y ago
      Insecticides in general are quite broad spectrum. You need to use different tactics entirely if you only want to target mosquitoes.
  - joshvm 6y ago
    In the UK there have been fairly successful movements to stop councils trimming verges, since it promotes wildflowers and insects. This was pre-covid, but probably it's helped. You should check out Wilding by the serendipitously named Isabella Tree, if you've not already.
Budabellly 6y ago
This discussion thread has me wondering if there's a good opportunity between the ubiquity of modern cameras (drones/satellites/phones) + computer vision to reduce the cost of those "total count" surveys you mentioned. Not that I was prepping for an interview but a couple questions I asked myself... * How could you use social media to establish the size of the pidgeon population in NYC? * Design a system for tracking Bluefin Tuna populations with an unlimited satellite imagery budget. * How many drones would it take to count the total number of wolves in Yellowstone in 1 week? * In safari destinations in Africa like the Maasai Mara, what % of megafauna are photographed annually? Makes you wonder if all of our facial recognition tech could find a better use in ecology.
- joshvm 6y ago
  This is basically what I've spent the last two years working on - my postdoc has been building an open-source drone payload that does real-time animal detection in visible/thermal images. Drones aren't practical for total counts. These tend to be e.g. an entire safari park. You need multiple planes flying for several hours each day. Huge expense - hundreds of thousands easily. What's more common is to do transect counts where you fly in a sort of raster pattern and then do your statistics to estimate the population that you didn't see. Drones are still massively limited by flight time, typically an hour at most. Maybe a bit more if you have a really good fixed wing. See https://www.nationalgeographic.com/news/2016/01/160104-great-elephant-census-vulcan-paul-allen-elephants-conservation/ https://www.nationalgeographic.com/news/2016/01/160104-great... On the other hand drones can be used as an effective anti-poaching deterrent - for example if you have a rough location. They can also be used for small scale surveys in locations where you can't really afford a plane and/or it's awkward or dangerous to send people. Some applications like flying very close to cliff faces to monitor birds are only possible to do safely with drones. Ecology has a whole host of problems that make object detection difficult. You have extreme class imbalance when you capture data - 90%+ of images are empty. Animals are frequently camouflaged. Many animals (like the various gazelles in Africa) look almost identical, especially in low resolution aerial imagery. A lot of animals of interest are small. Even for camera traps, where you have very controlled conditions, the standard approach these days is to have a generic "animal detector" and then have humans do the long tail classification. On the Africa question though - not a large percentage at all. By the way megafauna covers a lot more than it suggests (40kg+, so that includes deer). Most smaller conservancies have a good idea of how many "rare" species they have, because they're extremely conscious of poaching. But water buffalo? Thompsons gazelle? Who knows exactly. And once you go out of controlled areas into the bush proper, it's anyone's guess. There are charities which organise surveys, but it costs so much to do that it's not feasible to do it annually. See that Nat Geo link - that was the first time anyone had surveyed elephants in Angola. In 2016. But computer vision is absolutely the future. People are coming round to the idea that we need to start capturing full survey imagery during counts. And while those images are still labelled by people today, once we have that data stored we can always run new models over it in the future.
credit_guy 6y ago
I don't know anything about entomology or ecology, etc. But I noticed a stark decline in dead insects on the windshield of the car after a road trip in Europe over the last 20 years. Not so in the US. Meaning: 20 years ago I saw lots of dead insects on the windshield in Europe, very few in the US. Now it's very few in both places. So, I can corroborate: noticeable insect decline in Europe, no noticeable decline in the US, but unfortunately from a low level to begin with. Not sure if we are supposed to cheer this result.
- jrumbut 6y ago
  Funny you mention this. Having driven in the US, Mexico, and Italy (in recent years) I had never experienced a thing I had seen in movies, a windshield full of bugs. This was until I went to a remoter part of northern Quebec and found that without intervention the windshield would become opaque and driving with the window open created a horrifying mess near the rear window.
joshvm 6y ago
If anyone is interested in the paper I mentioned, you can read it here (open access): https://www.sciencedirect.com/science/article/pii/S0006320718318391 https://www.sciencedirect.com/science/article/pii/S000632071... I've worked with Richard and the folks at STE on a different digital imaging survey, they're doing some fantastic work and are really pioneers, despite a lot of push back from the community. You can also see some images in that paper which show how difficult this job is: https://ars.els-cdn.com/content/image/1-s2.0-S0006320718318391-gr6.jpg https://ars.els-cdn.com/content/image/1-s2.0-S00063207183183... Note different gazelle species - of course you'd have better resolution if you were looking in the flesh, but it's really a job for experts.

by joshvm 6y ago

DuskStar 6y ago

> but that relies on you knowing what those uncertainties are. And it also relies on those uncertainties not changing. "Elephants start spending more time in the forests and not open areas to avoid poachers" would reduce counts from airplanes without changing the number of elephants, for example.

Natsu 6y ago

For insects, there are a lot of different traps based on the nature of the insects being sought. There's a pretty good list here: https://en.wikipedia.org/wiki/Insect_trap https://en.wikipedia.org/wiki/Insect_trap As I understand the sampling, you pretty much just put the traps out and count up how many insects you get of each type. Purely anecdotally, I feel like the number of butterflies is way down from the level you'd see during, say, the 90s. It seems like I just don't see as many different types any more and even the more common types feel vastly reduced in number. I used to see monarchs and sulfers and painted ladies and red admirals and cabbage butterflies all the time, now I barely see a few cabbage butterflies or sulfers and occasionally a painted lady or two. I never get to see the less common things like buckeyes any more, which is a pity given that they're my favorite.

9nGQluzmnq3M 6y ago
This is anecdotal, but during Singapore's lockdown, one of the many services cut for about two months was curb maintenance. Being deep in the tropics, all green spaces immediately turned into jungles of flowers and weeds, and the butterfly population consequently exploded: https://mothership.sg/2020/06/grass-cutting-butterflies/ https://mothership.sg/2020/06/grass-cutting-butterflies/ https://www.bbc.com/news/world-asia-52960623 https://www.bbc.com/news/world-asia-52960623 On the other hand, mosquito populations also increased and there were lots of people stuck at home as bait for them, contributing to Singapore's worst dengue outbreak in years: https://www.tnp.sg/news/singapore/three-more-dengue-deaths-singapore-outbreak-worsens https://www.tnp.sg/news/singapore/three-more-dengue-deaths-s... So at least one lesson here is that, if given an opportunity to bounce back, many insect populations can and will.
- Natsu 6y ago
  Wow, Singapore has some really pretty ones. I've always been partial to butterflies with eyespots.
- refurb 6y ago
  Singapore is also going through a particularly bad Dengue fever season, so there is a ton being done for mosquito control including spraying of insecticides. I have no idea how selective those insecticides are, but one could imagine an impact on other insects.
  - CydeWeys 6y ago
    Insecticides in general are quite broad spectrum. You need to use different tactics entirely if you only want to target mosquitoes.
- joshvm 6y ago
  In the UK there have been fairly successful movements to stop councils trimming verges, since it promotes wildflowers and insects. This was pre-covid, but probably it's helped. You should check out Wilding by the serendipitously named Isabella Tree, if you've not already.

Budabellly 6y ago

This discussion thread has me wondering if there's a good opportunity between the ubiquity of modern cameras (drones/satellites/phones) + computer vision to reduce the cost of those "total count" surveys you mentioned. Not that I was prepping for an interview but a couple questions I asked myself... * How could you use social media to establish the size of the pidgeon population in NYC? * Design a system for tracking Bluefin Tuna populations with an unlimited satellite imagery budget. * How many drones would it take to count the total number of wolves in Yellowstone in 1 week? * In safari destinations in Africa like the Maasai Mara, what % of megafauna are photographed annually? Makes you wonder if all of our facial recognition tech could find a better use in ecology.

joshvm 6y ago
This is basically what I've spent the last two years working on - my postdoc has been building an open-source drone payload that does real-time animal detection in visible/thermal images. Drones aren't practical for total counts. These tend to be e.g. an entire safari park. You need multiple planes flying for several hours each day. Huge expense - hundreds of thousands easily. What's more common is to do transect counts where you fly in a sort of raster pattern and then do your statistics to estimate the population that you didn't see. Drones are still massively limited by flight time, typically an hour at most. Maybe a bit more if you have a really good fixed wing. See https://www.nationalgeographic.com/news/2016/01/160104-great-elephant-census-vulcan-paul-allen-elephants-conservation/ https://www.nationalgeographic.com/news/2016/01/160104-great... On the other hand drones can be used as an effective anti-poaching deterrent - for example if you have a rough location. They can also be used for small scale surveys in locations where you can't really afford a plane and/or it's awkward or dangerous to send people. Some applications like flying very close to cliff faces to monitor birds are only possible to do safely with drones. Ecology has a whole host of problems that make object detection difficult. You have extreme class imbalance when you capture data - 90%+ of images are empty. Animals are frequently camouflaged. Many animals (like the various gazelles in Africa) look almost identical, especially in low resolution aerial imagery. A lot of animals of interest are small. Even for camera traps, where you have very controlled conditions, the standard approach these days is to have a generic "animal detector" and then have humans do the long tail classification. On the Africa question though - not a large percentage at all. By the way megafauna covers a lot more than it suggests (40kg+, so that includes deer). Most smaller conservancies have a good idea of how many "rare" species they have, because they're extremely conscious of poaching. But water buffalo? Thompsons gazelle? Who knows exactly. And once you go out of controlled areas into the bush proper, it's anyone's guess. There are charities which organise surveys, but it costs so much to do that it's not feasible to do it annually. See that Nat Geo link - that was the first time anyone had surveyed elephants in Angola. In 2016. But computer vision is absolutely the future. People are coming round to the idea that we need to start capturing full survey imagery during counts. And while those images are still labelled by people today, once we have that data stored we can always run new models over it in the future.

credit_guy 6y ago

I don't know anything about entomology or ecology, etc. But I noticed a stark decline in dead insects on the windshield of the car after a road trip in Europe over the last 20 years. Not so in the US. Meaning: 20 years ago I saw lots of dead insects on the windshield in Europe, very few in the US. Now it's very few in both places. So, I can corroborate: noticeable insect decline in Europe, no noticeable decline in the US, but unfortunately from a low level to begin with. Not sure if we are supposed to cheer this result.

jrumbut 6y ago
Funny you mention this. Having driven in the US, Mexico, and Italy (in recent years) I had never experienced a thing I had seen in movies, a windshield full of bugs. This was until I went to a remoter part of northern Quebec and found that without intervention the windshield would become opaque and driving with the window open created a horrifying mess near the rear window.

joshvm 6y ago

If anyone is interested in the paper I mentioned, you can read it here (open access): https://www.sciencedirect.com/science/article/pii/S0006320718318391 https://www.sciencedirect.com/science/article/pii/S000632071... I've worked with Richard and the folks at STE on a different digital imaging survey, they're doing some fantastic work and are really pioneers, despite a lot of push back from the community. You can also see some images in that paper which show how difficult this job is: https://ars.els-cdn.com/content/image/1-s2.0-S0006320718318391-gr6.jpg https://ars.els-cdn.com/content/image/1-s2.0-S00063207183183... Note different gazelle species - of course you'd have better resolution if you were looking in the flesh, but it's really a job for experts.