LOCA version 1 vs. LOCA version 2

LOCA version 1 was used to downscale the CMIP5 data and is approximately vintage 2016.

LOCA version 2 was used to downscale the CMIP6 data and is approximately vintage 2023.

The main differences between LOCA version 1 and LOCA version 2 are:

  1. LOCA version 1 used Livneh et al. 2015 as the precipitation training data set. However after producing LOCA version 1 it became apparent that Livneh et al. 2015 depicts weaker precipitation extremes than are observed (e.g., Pierce et al. 2021, Risser et al. 2021). I created a new precipitation training data set that does a better job representing precipitation extremes (Pierce et al. 2021), which can be downloaded here. I used this new data set, which I call “unsplit Livneh”, as the precipitation training data set in LOCA version 2. The temperature data set in LOCA version 2 is still the Livneh 2015 data set, but extended to 2018 by Lu Su at UCLA. It is also available for download at that link above.
  2. In LOCA version 1 we downscaled one ensemble member per model. In LOCA version 2 we downscaled up to 10 ensemble members per model, if the model had multiple ensemble members available.
  3. In LOCA version 2, when a model had multiple ensemble members available, precipitation was bias corrected using the multi-ensemble member bias correction scheme described in Pierce et al. 2023. In brief, the data from all ensemble members is pooled and bias corrected together. This approach does a better job preserving rare, extreme precipitation events than you would get if each ensemble member were bias corrected individually.
  4. In LOCA version 1 (CMIP5), we downscaled 2 Representative Concentration Pathways (RCP’s): RCP 4.5 and RCP 8.5, which can generally be thought of as medium-low and high emissions scenarios, respectively. In LOCA version 2 (CMIP6), we downscaled 3 Shared Socioeconomic Pathways (SSPs): SSP 245, SSP 370, and SSP 585, which can generally be thought of as medium-low, medium, and high emissions scenarios, respectively. SSP 245 is similar to RCP 4.5, and SSP 585 is similar to RCP 8.5. SSP 370 is a “medium” emissions scenario, and is similar to the trajectory the Earth is currently experiencing.