DSEclimatekitaeExt

Uses Cal-Adapt’s climakitae, which does do a good job of abstracting a lot of the fetching, processing, and visualize downscaled CMIP6 climate data for most of the western US (with best resolutions supported only for california and very adjacent spaces).

Features¶

• Distributed data fetching — Sends work to Coiled cloud workers that sit next to the data on S3. Workers open Zarr stores, subset by time and space, convert units, and return results. You don’t download raw data to your laptop.

• Powerful helper functions - Instead of owning work across the whole stack, some helper functions only request that you know variables, time scales, scenarios and parks you’d like data from.

• De(Myst)ification - Actually don’t even know what park you want? Or your wish of variable depends on what resolution and scenarios are available? Helper functions send live probes into CalAdapts zarr stores to confidently tell you what’s available for where at resolution and over what time period.

• Fuzzy park search — ParkCatalog searches across all 437 NPS units. Type “yosemite” or even “grand canion” and it finds the right park. No need to know exact names.

• Full scenario coverage — Historical Climate (1950-2014) plus three future scenarios (SSP 2-4.5, SSP 3-7.0, SSP 5-8.5) across 15 CMIP6 global climate models at 3km monthly resolution (LOCA2 statistical downscaling).

• Spatial heatmaps — 2x2 comparison plots showing how temperature or precipitation patterns change across scenarios within park boundaries. Anomaly plots show the difference from the historical baseline.

• CSV export — Clean CSVs with consistent column names (simulation, time, scenario, T_Max, etc.) that read straight into R with read.csv() or readr::read_csv().

Get Started¶

Need Setup walkthrough? Easiest to just start with with the installation guide to get the container running, then open the Tutorial notebook.

• Installation — Set up Docker + VS Code

• Getting Started Overview — What to do first

Tutorials¶

• Tutorial Walkthrough

• ParkCatalog Demo

• Full Data Extraction

• Spatial Comparison

Getting Started with External Data¶

:P

Project Structure¶

:P

Advanced¶

• Pipeline (Luigi)

• Formal Python packaging

• WRF dynamical downscaling

• Setting up coiled for your own team

Links¶

• GitHub Repository

• Cal-Adapt Analytics Engine

• climakitae

• Coiled

• LOCA2 Downscaling