About Me…

Thanks also to …

About You

Number of registrants: 58

Type of Organization

UC Campus

Familiarity with Climate Change Science

Familiarity with R

Location

Outcomes

1) Get familiar with caladaptR

2) Hands-on practice with:

1. importing Cal-Adapt into R via the API
   
2. data wrangling techniques to prep data for visualization or analysis
   
3. see techniques for dealing with large volumes of data

3) Working Code Recipes

Cal-Adapt

Cal-Adapt is California’s official portal for peer-reviewed climate data.

Datasets are selected with guidance and priorities from California State agencies.

Modeled Climate Data:

See also: What climate data does Cal-Adapt provide?.

Spatial extent of LOCA downscaled climate data layers:

Climate Data

Where are the Data?

Options for Accessing Cal-Adapt Data

More info:

Why you might want to work with Cal-Adapt data in R

caladaptR

Key Features

Main Uses

Prerequisites

caladaptR users need to know:

caladaptR workflow

In general, there are five steps to using caladaptR:

Building an API Request

cap1 <- ca_loc_pt(coords = c(-121.4687, 38.5938)) %>%  
  ca_cvar(c("tasmax", "tasmin")) %>%                   
  ca_gcm(gcms[1:4]) %>%                                
  ca_scenario(scenarios[1:2]) %>%                      
  ca_period("year") %>%                                
  ca_years(start = 2040, end = 2060)                   

An API Request object consists of between 2 and 4 components:

1. Location (required, pick one)

2. Raster series dataset

3. Start & end dates (optional, pick one)

4. Options (only required for polygons)

Quick Example

1. Load the package:

library(caladaptr)

2. Create an API request object:

cap1 <- ca_loc_pt(coords = c(-121.4687, 38.5938)) %>%  ## specify a location
  ca_cvar(c("tasmax", "tasmin")) %>%                   ## climate variables
  ca_gcm(gcms[1:4]) %>%                                ## GCM(s)
  ca_scenario(scenarios[1:2]) %>%                      ## emission scenarios(s)
  ca_period("year") %>%                                ## temporal aggregation period
  ca_years(start = 2040, end = 2060)                   ## start and end dates

3. Check API request

cap1

## Cal-Adapt API Request
## Location(s): 
##   x: -121.469
##   y: 38.594
## Variable(s): tasmax, tasmin
## Temporal aggregration period(s): year
## GCM(s): HadGEM2-ES, CNRM-CM5, CanESM2, MIROC5
## Scenario(s): rcp45, rcp85
## Dates: 2040-01-01 to 2060-12-31
## 

cap1 %>% ca_preflight()

## General issues

##  - none found

## Issues for querying values

##  - none found

## Issues for downloading rasters

##  - none found

plot(cap1, locagrid = TRUE)

4. Fetch data:

cap1_tbl <- ca_getvals_tbl(cap1, quiet = TRUE)

5. View the results

cap1_tbl %>% slice(1:10)

R Notebooks

R Notebooks are written in “R Markdown”, which combines text and R code.

Notebook 1. Getting Started

In Notebook 1, you will:

Notebook 1. Getting Started | solutions

Other Ways to Specify Location: Preset Areas-of-Interest

The Cal-Adapt API server has 11 preset areas of interest (aka boundary layers).

You can query these features without passing a spatial object!

Example:

norcal_cap <- ca_loc_aoipreset(type = "counties", 
                         idfld = "fips", 
                         idval = c("06105", "06049", "06089", "06015")) %>% 
  ca_gcm(gcms[1:4]) %>%
  ca_scenario(scenarios[1:2]) %>%
  ca_period("year") %>%
  ca_years(start = 2040, end = 2060) %>%
  ca_cvar("pr") %>% 
  ca_options(spatial_ag = "max")

Plot an API request to verify the location:

plot(norcal_cap)

Other Ways to Specify Location: User-Defined Polygon

An API request can use a simple feature data frame as the query location (point, polygon, and multipolygon).

Use sf::st_read() to import Shapefiles, geojson, KML, geopackage, ESRI geodatabases, etc.

library(sf)
pinnacles_bnd_sf <- st_read("./notebooks/data/pinnacles_bnd.geojson")

## Reading layer `pinnacles_bnd' from data source 
##   `D:\GitHub\cal-adapt\caladaptr-res\docs\workshops\ca_intro_feb22\notebooks\data\pinnacles_bnd.geojson' 
##   using driver `GeoJSON'
## Simple feature collection with 1 feature and 5 fields
## Geometry type: MULTIPOLYGON
## Dimension:     XY
## Bounding box:  xmin: -121.2455 ymin: 36.4084 xmax: -121.1012 ymax: 36.56416
## Geodetic CRS:  WGS 84

pinnacles_bnd_sf

Crate an API request:

pin_cap <- ca_loc_sf(loc = pinnacles_bnd_sf, idfld = "UNIT_CODE") %>% 
  ca_slug("et_month_MIROC5_rcp45") %>%
  ca_years(start = 2040, end = 2060) %>% 
  ca_options(spatial_ag = "mean")

plot(pin_cap, locagrid = TRUE)

Fetch data:

pin_tbl <- pin_cap %>% ca_getvals_tbl()
head(pin_tbl)

Using Climate Data Wisely

Which weekend in February 2070 has the least chance of rain for my retirement party at the San Luis Obispo Country Club?

Wise or unwise?

library(caladaptr)
library(dplyr)
library(ggplot2)
library(units)

slo_cap <- ca_loc_pt(coords = c(-120.6276, 35.2130)) %>%
  ca_cvar("pr") %>%
  ca_gcm(gcms[1:4]) %>%
  ca_scenario("rcp45") %>%
  ca_period("day") %>%
  ca_dates(start = "2070-02-01", end = "2070-02-28")

slo_mmday_tbl <- slo_cap %>% 
  ca_getvals_tbl(quiet = TRUE) %>% 
  mutate(pr_mmday = set_units(as.numeric(val) * 86400, mm/day))

feb2070_weekends_df <- data.frame(dt = as.Date("2070-02-01") + c(0,1,7,8,14,15,21,22), y = 50)

ggplot(data = slo_mmday_tbl, aes(x = as.Date(dt), y = as.numeric(pr_mmday))) +
  geom_col(data = feb2070_weekends_df, aes(x = dt, y = y), fill = "red", alpha = 0.2) +
  geom_point(aes(color=gcm)) + 
  labs(title = "Chance of Rain at the San Luis Obispo Country Club, February 2070", 
       subtitle = "Pink bars represent the weekends. Emissions scenario: RCP 4.5.",
       x = "date", y = "mm") +
  theme(plot.caption = element_text(hjust = 0)) 

1) Examine 20-30 Year Time Periods

2) Look at multiple models & emissions scenarios

3) Aggregate, aggregate, aggregate

4) Look at both central tendency and the range of variability

Both are important.

Variability can be within and/or between models.

4) Compare, compare, compare

Raster Series Catalog

caladaptR comes with a Raster Series Catalog

For each raster series the catalog will list its:

The catalog can be retrieved using ca_catalog_rs() (returns a tibble).

The best way to browse the catalog is with RStudio’s View pane. You can then use the filter buttons to find the raster series you want.

View(ca_catalog_rs())

Data Wrangling

Cal-Adapt data always comes down in a “long” format:

head(cap1_tbl)

To make useful plots, maps, and summaries, you often have to ‘munge’ or ‘wrangle’ the data, which may include:

Fortunately you have a very robust toolbox:

cheatsheet

cheatsheet

cheatsheet

cheatsheet

Useful functions

Qtn. How do I know what data wrangling is needed?

Ans. Work backward from your analysis or visualization functions.

Notebook 2. Find and Wrangle Data

In Notebook 2, you will:

Notebook 2. Finding Data | solutions

Shiny Apps

• The Shiny package allows you to create browser-based apps that run R in the background.

• You need to know to R, but you don’t need to know JavaScript, HTML, CSS, network protocols, etc.

• Can run them locally or online

library(caladaptr.apps)

## Launch the time series app
ca_launch("timeseries")

## Launch the projected chill portions app
ca_launch("chill")

Notebook 3. Shiny Apps

Notebook 3. Shiny Apps

Additional Resources

https://ucanr-igis.github.io/caladaptr/

Work in progress…

Test Cases and Developers Wanted!

END!