> A new first-of-its-kind study by Dutch researchers
reading through [0] the affiliations of the "researchers" is the first red flag I notice.
the lead author is Hessel Voortman, from the Hessel Voortman Engineering Consultancy. his co-author is Rob De Vos, an "independent researcher".
by itself, that isn't necessarily a problem...but I'd prefer to read papers about climate change from actual experts in climate change, not Guy Who Owns Engineering Company and His Buddy
the next problem, from the abstract:
> We used two datasets with local sea level information all over the globe. In both datasets, we found approximately 15% of the available sets suitable to establish the rate of rise in 2020.
only 15% of the data being suitable seems kind of surprising. they elaborate on their criteria:
> The present study aims to estimate the long-term rates of sea level rise in 2020. To do so reliably, data were selected according to the following criteria:
> - Latest year in the dataset not earlier than 2015, this being a compromise between the desire to have data up to 2020 and the desire to have as many locations included as possible
> - Data available over a period of at least 60 years
> - At least 80% of the years in the range with data available
> Satellite data do not fulfill our second criterion and were, therefore, not used.
so they have a bunch of data, they come up with some cutoffs that discard 85% of the data, then draw conclusions from the other 15%?
in particular, satellite data hasn't been around for 60 years, so...toss it all out, I guess? couldn't satellite data from the past 30 years still be used to support their findings? couldn't every-other-year data fill the same purpose as well?
this smells a bit like p-hacking...if you can exclude data based on arbitrary criteria, you can reach any conclusion you want by excluding the right data.
but anyways, I think it's heartwarming that climate denial is still being done by humans. you'd think AI would have taken over that job by now.
reading through [0] the affiliations of the "researchers" is the first red flag I notice.
the lead author is Hessel Voortman, from the Hessel Voortman Engineering Consultancy. his co-author is Rob De Vos, an "independent researcher".
by itself, that isn't necessarily a problem...but I'd prefer to read papers about climate change from actual experts in climate change, not Guy Who Owns Engineering Company and His Buddy
the next problem, from the abstract:
> We used two datasets with local sea level information all over the globe. In both datasets, we found approximately 15% of the available sets suitable to establish the rate of rise in 2020.
only 15% of the data being suitable seems kind of surprising. they elaborate on their criteria:
> The present study aims to estimate the long-term rates of sea level rise in 2020. To do so reliably, data were selected according to the following criteria:
> - Latest year in the dataset not earlier than 2015, this being a compromise between the desire to have data up to 2020 and the desire to have as many locations included as possible
> - Data available over a period of at least 60 years
> - At least 80% of the years in the range with data available
> Satellite data do not fulfill our second criterion and were, therefore, not used.
so they have a bunch of data, they come up with some cutoffs that discard 85% of the data, then draw conclusions from the other 15%?
in particular, satellite data hasn't been around for 60 years, so...toss it all out, I guess? couldn't satellite data from the past 30 years still be used to support their findings? couldn't every-other-year data fill the same purpose as well?
this smells a bit like p-hacking...if you can exclude data based on arbitrary criteria, you can reach any conclusion you want by excluding the right data.
but anyways, I think it's heartwarming that climate denial is still being done by humans. you'd think AI would have taken over that job by now.
0: https://www.mdpi.com/2077-1312/13/9/1641