You seem very angry about this.
Here's what we know:
The paper compares temperature data primarily from a single weather station 'El Verde', starting in 1978. [1]
The operators of the weather station say themselves that the data is unreliable, that papers have already drawn erroneous conclusions based on it, and that it shouldn't be used for calculating long term trends. [2]
Here's a few quotes from [2], but please read the whole thing.
"Subsequent manipulations make the EVFS record less valuable for interpreting long term trends."
"After September 1992, recorded maximum temperatures increased substantially at this station, certainly as a result of the instrumentation changes that took place in that month."
"correction ceased to be applied to these maximum temperature data in 1997, and as a result, the EVFS maximum temperature record (LUQ data set 16) showed an abrupt increase in maximum temperature in 1997."
"This partially corrected data set is the one we made accessible to people on our web site and through ClimDB until Feb 2014. At least one paper has reached erroneous conclusions about long term trends based on these data (Huey et al. 2009)."
"The differences in instrumentation have introduced differences in the data between these two periods, and treating them as a single data set misrepresents the data."
Knowing this, I think it's fair to say that any data prior to 1997 is invalid. Now take the graph in [1] and disregard 1997 backwards, and it paints an entirely different picture doesn't it?
[1] https://notalotofpeopleknowthat.files.wordpress.com/2019/01/...
[2] http://luq.lternet.edu/data/luqmetadata16