Time Series Analysis

Max Anjos

March 25, 2026

Getting Started

The lcz_ts() function allows you to analyze air temperature data associated with Local Climate Zones (LCZ) over time. In this tutorial, we’ll perform an hourly frequency analysis of air temperature in Berlin for the years 2019-2020 using lcz_data from the LCZ4r package.

library(LCZ4r)

# Get the LCZ map for your city
lcz_map <- lcz_get_map_euro(city = "Berlin")

# Load sample data from LCZ4r
data("lcz_data")

Basic Time Series Analysis

Let’s start with a basic analysis using lcz_ts() to observe hourly temperature fluctuations.

# Get monthly LCZ-air temperature time series
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station",
       time.freq = "month",
       ylab = "Air temperature [°C]",
       xlab = "Date",
       title = "LCZ - Time Series",
       caption = "Source: LCZ4r, 2024.")

Monthly LCZ-air temperature time series for Berlin showing seasonal temperature variations across different LCZ classes

Plotting Options with plot_type

The plot_type argument in lcz_ts() offers several visualizations:

• “basic_line”: Simple line plot
• “faceted_line”: Line plot with facets by LCZ or station
• “heatmap”: Heatmap to visualize temporal patterns
• “warming_stripes”: Stripes representing temperature variations over time

Below are examples using each plot type.

1. Basic Line Plot

# Daily air temperature for September 2019
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station",
       time.freq = "day",
       year = 2019, month = 9,
       plot_type = "basic_line")

Basic line plot showing daily temperature variations in Berlin during September 2019

2. Faceted Line Plot by Station

# Daily air temperature from January to March 2019, faceted by station
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "hour",
       year = 2019, month = 1:3,
       plot_type = "facet_line", 
       facet = "station")

Faceted line plot showing temperature patterns across different meteorological stations in Berlin (January-March 2019)

3. Heatmap of Temperature Over Time

# Hourly air temperature on January 15, 2020, with LCZ facets
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "hour",
       year = 2020, month = 1, day = 15,
       plot_type = "heatmap", 
       facet = "LCZ")

Heatmap showing hourly temperature variations across different LCZ classes on January 15, 2020

4. Warming Stripes

# Hourly air temperature for 2020, visualized as warming stripes
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "hour", 
       year = 2020, month = 1:12,
       plot_type = "warming_stripes",
       facet = "LCZ")

Warming stripes visualization showing temperature trends across different LCZ classes throughout 2020

Splitting Data with the “by” Argument

You can split data by temporal or spatial categories such as “year”, “month”, “season”, “weekday”, “weekend”, “yearseason”, and more.

# Daily air temperature by seasons of 2020
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "day",
       year = 2020,
       plot_type = "basic_line", 
       by = "season")

Temperature variations across different seasons in Berlin during 2020

Dividing Nighttime and Daytime with “by” Argument

You can also split data by nighttime and daytime periods with by = "daylight".

# Diurnal cycle of air temperature on January 15, 2020
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "hour",
       year = 2020, month = 1, day = 15,
       plot_type = "heatmap", 
       by = "daylight")

Comparison of daytime and nighttime temperature patterns on January 15, 2020

Combining Daylight with Months

# Diurnal cycle of air temperature on January 15 and July 15, 2020
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "hour",
       year = 2020, month = c(1, 7), day = 15,
       plot_type = "basic_line", 
       by = c("daylight", "monthyear"))

Comparison of diurnal temperature cycles in winter (January) and summer (July) 2020

Smoothing Temperature Trends

Enable smoothing by setting smooth = TRUE, which adds a generalized additive model (GAM) line to show the temperature trend.

# Trends of daily air temperature for 2019-2020 by station
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "hour",
       year = 2019:2020,
       plot_type = "basic_line", 
       smooth = TRUE)

Temperature trends with GAM smoothing across stations for 2019-2020

Save Plots

To save a plot and dataframe, set isave = TRUE and specify the file type with save_extension (e.g., “png”, “jpeg”, “svg”, “pdf”). A folder LCZ4r_output is created on your PC.

# Save daylight plot and dataframe to PC
lcz_ts(lcz_map, 
       data_frame = lcz_data, 
       var = "airT", 
       station_id = "station", 
       time.freq = "hour",
       year = 2020, month = 1, day = 15,
       plot_type = "basic_line", 
       by = "daylight", 
       isave = TRUE)

Tip: The saved files will be automatically organized in the LCZ4r_output folder with timestamped filenames for easy reference.

Return a Dataframe as Result

To save the result in R, set iplot = FALSE and create an object.

# Return daylight dataframe for January 15, 2020
my_output <- lcz_ts(lcz_map, 
                    data_frame = lcz_data, 
                    var = "airT", 
                    station_id = "station", 
                    time.freq = "hour",
                    year = 2020, month = 1, day = 15,
                    plot_type = "basic_line", 
                    by = "daylight", 
                    iplot = FALSE)

# View the structure of the returned dataframe
str(my_output)

Summary of Parameters

Here’s a quick reference for the main parameters used in lcz_ts():

Parameter	Description	Options
time.freq	Temporal aggregation frequency	“hour”, “day”, “month”, “year”
plot_type	Visualization type	“basic_line”, “facet_line”, “heatmap”, “warming_stripes”
by	Data splitting method	“season”, “month”, “weekday”, “daylight”, “yearseason”
smooth	Add GAM trend line	TRUE/FALSE
isave	Save output to PC	TRUE/FALSE
iplot	Display plot	TRUE/FALSE

---

Have feedback or suggestions?

Do you have an idea for improvement or did you spot a mistake? We’d love to hear from you! Click the button below to create a new issue (GitHub) and share your feedback or suggestions directly with us.

Open GitHub issue