jinshijian commented 4 years ago

Done this week

Surrogates

Prepared an outline and shared with coauthors
Kalyn's question?
Is it worth showing spatial maps (or differences between data products) in the SI to understand anything about the observational uncertainty? How?
Rh latitude pattern: why using % of global, why Hashimoto has a relatively large value at Latitude ~85, because other dataset does not have a prediction at this region?

Kick off the "Time for warm" project?
Reading papers recommended by @bpbond and Dalei, thinking about the broad picture of the landmodels project, and how to tell the store.

jinshijian commented 4 years ago

Chris: Would the analysis be cleaner, with biases more apparent if you evaluated instantaneous values rather than integrated and scaled values for the year?
Jinshi: maybe we can use data from COSORE to do an additional test!
Measure interval: also using data from COSORE to do the test?

jinshijian commented 4 years ago

Temperature increase in all regions in the past 50 years;
And air temperature of 1990s (air temperature measure before 2000) is lower than that from 2000s (air temperature measured after 2000);
I collected 48 studies, each study have measured annual Rs more (or =) 5 years of Rs, and some studies has multiple sites (or treatment), totally I have 84 samples (for the first 5 years);
One study has 26 years of Rs_annual records;
Then, each study, I calculated the temperature and Rs_annual anomaly;
The temperature anomaly is as expected (increase trend because of global warming, see the figure panel a);
However, the Rs anomaly showed a decreasing trend! (see the figure panel b)
Why?
It is very surprising to find this result!