There's no research going on, because this isn't something you can address scientifically with any semblance of hope of getting it.
So what we have instead is modelling, aka "data science". It's essentially numerology pretending to use tools of science.
Why do I say that?
Science is about a specific process: you make a hypothesis, you set up a test of your hypothesis, you test it, find it true or not and based on that your hypothesis becomes a scientific theory or a rejected hypothesis.
"Data science" works in exact opposite way. It generates no hypothesis, instead if generates "how world should work according to our best guess" model. I.e. it starts with the outcome "this is how world should be, now let us pick the numbers to demonstrate it", rather than the premise "world may be this way, let's see if it's true". Then it proceeds to plug numbers, both guessed and real into the model. Then they see if the output, and begin adjusting the numbers to get the outputs they desire.
As a result, science doesn't care about scientist. Hypothesis is either repeatable or not. Modelling is all about the human. Which numbers have been selected, which left out, which numbers have been guessed, and what multipliers were assigned.
And that is why "environmental modelling" has been so hilariously wrong all the time. Remember all the "2020 is the time when all the models from 2000s claiming to predict what happens in 20 years had to be actually tested against reality, and they all ran hilariously hot compared to reality". That was IPCC's reckoning at the time.
Nothing meaningful has changed since then. Modelling is still "top tier climate science".