Global Sea Surface Temperature (SST) data

Global Sea Surface Temperature (SST) data#

In this tutorial, we will look at how to access and analyse global sea surface temperature (SST) data.

Included:

Presentation and description of different sea surface temperature datasets, including their differences and the scenarios in which each is used.
A demonstration of how to access and analyse the data using Python.
- Level-3 data
- Level-4 data

What Do We Mean by Sea Surface Temperature?#

At first glance, defining sea surface temperature (SST) might seem straightforward, but the ocean’s temperature can vary significantly within its uppermost layers due to factors like weather conditions and ocean currents. Therefore, it is essential to specify the depth range being measured. This depends on the instrument used and the wavelength band measured.

Infrared radiometers measure the skin temperature of the ocean, usually sensing the uppermost 10 µm (micrometres) of the surface.
Microwave radiometers measure the subskin temperature, penetrating deeper; approximately 1 mm below the surface.

Consequently, SST measurements from different instruments may not always align.

SST depth profile

Figure from Group for High Resolution Sea Surface Temperature (GHRSST)

Measuring Sea Surface Temperature from Space#

Satellites measure SST through passive remote sensing, meaning they do not emit signals but instead detect naturally occurring radiation emitted or reflected. However, SST is not measured directly; rather, it is inferred through a series of calibration and correction steps applied to the raw data collected by the satellite’s radiometer.

SST is typically measured using infrared radiometers or microwave radiometers, each with its own advantages and limitations.

Infrared Radiometers#

Infrared radiometers measure thermal infrared radiation emitted by the ocean’s surface, providing high-resolution SST estimates under clear-sky conditions.

Examples

Advanced Very High Resolution Radiometer (AVHRR): https://www.class.noaa.gov/data_available/avhrr/index.htm
Sentinel-3 Sea and Land Surface Temperature Radiometer (SLSTR): https://sentiwiki.copernicus.eu/web/s3-slstr-instrument

Advantages
✔ High spatial resolution, enabling detailed SST maps.
✔ Accurate SST retrievals under cloud-free conditions.

Limitations
✘ Cannot measure SST through clouds, as water vapor absorbs infrared radiation.
✘ More affected by atmospheric interference, requiring more correction algorithms.

Microwave Radiometers#

Microwave radiometers measure naturally emitted microwave radiation from the ocean surface, providing SST estimates even in cloudy conditions. Microwave emissions penetrate atmospheric water vapor more effectively than infrared radiation.

Examples

GCOM-W Advanced Microwave Scanning Radiometer 2 (AMSR2): https://www.earthdata.nasa.gov/data/instruments/amsr2
SMOS Microwave Imaging Radiometer using Aperture Synthesis (MIRAS): https://earth.esa.int/eogateway/instruments/miras
TROPICS (NASA)

Advantages
✔ Can measure SST through clouds, making it ideal for all-weather monitoring.
✔ Less sensitive to atmospheric interference than infrared methods.

Limitations
✘ Lower spatial resolution, meaning SST estimates are less detailed.
✘ More affected by surface roughness (e.g., wind-driven waves).

Data Processing Levels#

SST products exist at different processing levels, from raw instrument readings to fully interpolated datasets. In this tutorial we will show you how to acces data from Level-3 and Level-4 data that are useful for most potential data users.

Level-2: Native Resolution SST Products#

Satellite radiometers collect raw digital data (Level-0) that is subsequently calibrated and converted into brightness temperatures (Level-1). These brightness temperatures are then processed—with atmospheric corrections applied—to derive Sea Surface Temperature (SST) products on the satellite’s native grid (Level-2). Data are served in one file per satellite image. May require additional processing or interpolation before integration into larger-scale studies.

When to use:

When you need high-resolution data with exact observation times.
For studying local or short-term events.
If you plan to apply your own processing or interpolation.

Sentinel-3 Level-2 products

Map showing outlines of Sentinel-3 Level-2 products.

Level-3: Gridded SST Products#

The data are mapped onto a regular spatial grid at various resolutions (e.g., daily or twice daily). The data are served in global or regional products. Although some of the native resolution detail is lost during the gridding process, Level-3 data strike a balance between spatial detail and ease of use.

When to use:

For relatively high-resolution regional or global studies.
When you need data on a standard grid for easier integration and analysis.

Level-3 data: OSI-201-b Level-3 data: Ocean and Sea Ice Satellite Application Facility (OSI-SAF) OSI-201-b

Level-4: Gap-Filled and Model-Interpolated SST#

Finally, Level-4 SST products incorporate provide gap-free data, sometimes combining Level-3 satellite observations with in-situ measurements (e.g., from buoys and ships) and model-generated data. The fusion of multiple data sources mitigates issues such as cloud cover or sensor anomalies, ensuring consistent temporal and spatial coverage. However, users should note that the merging and interpolation processes may introduce some smoothing, which means that while Level-4 products are excellent for detecting long-term trends and patterns, they might not capture the finest-scale details present in the original satellite observations.

When to use:

When you require a continuous, gap-free dataset.
For long-term climate monitoring and trend analysis.
If you prefer a ready-to-use dataset even if it means some loss of the finer detail.

Level-4 data: NOAA 1/4° Daily Optimum Interpolation Sea Surface Temperature Level-4 data: National Ocean and Atmospheric Administration (NOAA) 1/4° Daily Optimum Interpolation Sea Surface Temperature (OISST)

Visual Comparison of Level-3 and Level-4 Data#

Level-3 vs Level-4 data A comparison of level-3 (left) vs level-4 (right) data. Data from 2024-06-01.

Preparing our Python session#

Before we begin accessing the data, let’s first import the modules that we will use today

import xarray as xr
import matplotlib.pyplot as plt 
import cartopy.crs as ccrs
from datetime import datetime
import cartopy.feature as cfeature
import plotly.graph_objects as go
import numpy as np

Level-3 data: OSI-201-b#

Level-3 data: OSI-201-b

The OSI-201-b product, known as the Global Metop Sea Surface Temperature (SST), is a Level 3C dataset produced by the Ocean and Sea Ice Satellite Application Facility (OSI SAF). It provides twice-daily global SST measurements derived from the Advanced Very High Resolution Radiometer (AVHRR) aboard the Metop-B satellite. These measurements are processed using a multispectral algorithm that utilises infrared channels at 3.7, 10.8, and 12.0 µm, along with a cloud mask to ensure data quality. Additional inputs from the European Centre for Medium-Range Weather Forecasts (ECMWF), including atmospheric temperature, humidity, and aerosol profiles, are incorporated to correct for regional and seasonal biases. The data is presented on a 0.05° regular grid in NetCDF4 format, compliant with the Group for High Resolution Sea Surface Temperature (GHRSST) Data Specification version 2. Regular validation is performed by comparing satellite estimates against buoy measurements, achieving a target accuracy with a monthly absolute mean difference of 0.5 K and a standard deviation of 0.8 K. This product is essential for operational meteorological services, oceanographic analyses, and environmental monitoring.

Data characteristics#

Spatial resolution: 0.05° latitude × 0.05° longitude.
Temporal coverage: 2 products per day (00:00, 12:00) since 2016
Includes: Sea surface temperatures, sea ice concentration, wind speed.

Useful Links#

Dataset Information Page
https://osi-saf.eumetsat.int/products/osi-201-b
THREDDS Catalogue
- Human interface: https://tds0.ifremer.fr/thredds/catalog/OSI-201-b-metop_b/catalog.html
- Machine interface: https://tds0.ifremer.fr/thredds/catalog/OSI-201-b-metop_b/catalog.xml

Crediting the Data Providers#

When using this dataset in publications or presentations, please provide the following citation:

OSI SAF Global Sea Surface Temperature (Metop-B/AVHRR) 2016-onwards, OSI-201-b, doi:10.15770/EUM_SAF_OSI_NRT_2011. EUMETSAT Ocean and Sea Ice Satellite Application Facility. Data extracted from [distributor/distribution mean]: ([extracted period],) ([extracted domains],) accessed [download date]

The content within square brackets “[…]” is to be replaced with whatever distributor, period, domain or download date is relevant to the data being used. The “extracted period” and “extracted domains” are only needed if a subset of the data is being used. Suggestions for “distributor/distribution mean”: OSI SAF FTP server/the EUMETSAT Data Centre/EUMETCast/name and URL of a redistributor…

Writing the data to a CSV file#

You can write all or a subset of the data to a pandas dataframe that you can export as a CSV file.

df = xrds['sea_surface_temperature'].to_dataframe()
df.head()

			sea_surface_temperature
time	lat	lon
2024-02-01 12:00:00	-89.974609	-179.974609	NaN
		-179.924805	NaN
		-179.875000	NaN
		-179.825195	NaN
		-179.775391	NaN

Or for a list of variables

df = xrds[['sea_surface_temperature', 'wind_speed', 'sea_ice_fraction']].to_dataframe()
df.head()

			sea_surface_temperature	wind_speed	sea_ice_fraction
time	lat	lon
2024-02-01 12:00:00	-89.974609	-179.974609	NaN	NaN	NaN
		-179.924805	NaN	NaN	NaN
		-179.875000	NaN	NaN	NaN
		-179.825195	NaN	NaN	NaN
		-179.775391	NaN	NaN	NaN

To write the dataframe to a CSV file:

#df.to_csv('global_sea_surface_temperatures.csv')

Converting from Kelvin to Celsius#

The sea_surface_temperature values are stored in degrees kelvin.

xrds['sea_surface_temperature'].attrs['units']

'kelvin'

We can convert the data to celsius.

sst_celsius = xrds['sea_surface_temperature'] - 273.15

Plotting the data for a day of our choosing#

Let’s use the code we used above to first isolate the data for a single day, then let’s build on that to quickly plot the data.

desired_date = '20240201'  # YYYYMMDD
desired_time = '120000'  # hhmmss - must be 000000 or 120000

date_obj = datetime.strptime(desired_date, "%Y%m%d")
year = date_obj.year
day_of_year = f"{date_obj.timetuple().tm_yday:03d}"  # Ensure 3-digit format

url = f'https://tds0.ifremer.fr/thredds/dodsC/OSI-201-b-metop_b/{year}/{day_of_year}/{desired_date}{desired_time}-OSISAF-L3C_GHRSST-SSTsubskin-AVHRR_SST_METOP_B_GLB-sstglb_metop01_{desired_date}_{desired_time}-v02.0-fv01.0.nc'
xrds = xr.open_dataset(url)

sst_celsius = xrds['sea_surface_temperature'] - 273.15

sst_celsius.plot()
plt.show()

Without the coastlines these data are difficult to interpret. Let’s add them, as well as different colours for the land and sea.

We should also choose a map projection.

Map projections: https://scitools.org.uk/cartopy/docs/v0.15/crs/projections.html
Matplotlib built-in colourmaps: https://matplotlib.org/stable/users/explain/colors/colormaps.html
Colourmaps for oceanography from cmocean: https://matplotlib.org/cmocean/

# Loading in the data
desired_date = '20240601'  # YYYYMMDD
desired_time = '120000'  # hhmmss - must be 000000 or 120000

date_obj = datetime.strptime(desired_date, "%Y%m%d")
year = date_obj.year
day_of_year = f"{date_obj.timetuple().tm_yday:03d}"  # Ensure 3-digit format

url = f'https://tds0.ifremer.fr/thredds/dodsC/OSI-201-b-metop_b/{year}/{day_of_year}/{desired_date}{desired_time}-OSISAF-L3C_GHRSST-SSTsubskin-AVHRR_SST_METOP_B_GLB-sstglb_metop01_{desired_date}_{desired_time}-v02.0-fv01.0.nc'
xrds = xr.open_dataset(url)

# Setting up the figure
projection = ccrs.Mollweide()
transform = ccrs.PlateCarree()

fig = plt.figure(figsize=(16, 8))
ax = plt.axes(projection=projection)

ax.add_feature(cfeature.OCEAN, zorder=0, facecolor='black')
ax.add_feature(cfeature.LAND, zorder=0, facecolor='lightgrey')
ax.gridlines(draw_labels=True)
ax.coastlines()

# Plotting the data
sst_celsius = xrds['sea_surface_temperature'] - 273.15
vmin = -2
vmax = sst_celsius.max()
sst_celsius.plot(ax=ax, transform=transform, cmap='jet', vmin=vmin, vmax=vmax)
plt.savefig('sst_level3.png', transparent=True)
plt.show()

_images/47d60bf0cffce06586a665a9d00fa4ac286d2d61b76f40b51ab8711e5fdaa982.png

Finding the sea surface temperature at a given location and date#

Let’s now look at how to find the value for the sea surface temperature at a given location and time. We will using the method nearest to find the nearest value to our selected point, though other methods are available. https://docs.xarray.dev/en/latest/generated/xarray.Dataset.sel.html

desired_date = '20240201'  # YYYYMMDD
desired_time = '120000'  # hhmmss - must be 000000 or 120000
desired_latitude = 40
desired_longitude = -12

date_obj = datetime.strptime(desired_date, "%Y%m%d")
year = date_obj.year
day_of_year = f"{date_obj.timetuple().tm_yday:03d}"  # Ensure 3-digit format

url = f'https://tds0.ifremer.fr/thredds/dodsC/OSI-201-b-metop_b/{year}/{day_of_year}/{desired_date}{desired_time}-OSISAF-L3C_GHRSST-SSTsubskin-AVHRR_SST_METOP_B_GLB-sstglb_metop01_{desired_date}_{desired_time}-v02.0-fv01.0.nc'
xrds = xr.open_dataset(url)

sst_value = xrds.sel(
    time = desired_date,
    lat = desired_latitude,
    lon = desired_longitude,
    method = 'nearest'
)['sea_surface_temperature'].values

print(sst_value, xrds['sea_surface_temperature'].attrs['units'])

289.97999999999996 kelvin

We might also want to find out what latitude, longitude and time that value corresponds to.

desired_date = '20240201'  # YYYYMMDD
desired_time = '120000'  # hhmmss - must be 000000 or 120000

date_obj = datetime.strptime(desired_date, "%Y%m%d")
year = date_obj.year
day_of_year = f"{date_obj.timetuple().tm_yday:03d}"  # Ensure 3-digit format

url = f'https://tds0.ifremer.fr/thredds/dodsC/OSI-201-b-metop_b/{year}/{day_of_year}/{desired_date}{desired_time}-OSISAF-L3C_GHRSST-SSTsubskin-AVHRR_SST_METOP_B_GLB-sstglb_metop01_{desired_date}_{desired_time}-v02.0-fv01.0.nc'
xrds = xr.open_dataset(url)

single_point = xrds.sel(
    lat = desired_latitude,
    lon = desired_longitude,
    method = 'nearest'
)

latitude_value = single_point['lat'].values
longitude_value = single_point['lon'].values
sst_value = single_point['sea_surface_temperature'].values

print('Latitude:',latitude_value)
print('Longitude:',longitude_value)
print('Sea surface temperature value:',sst_value, xrds['sea_surface_temperature'].attrs['units'])

Latitude: 40.02539
Longitude: -11.974609
Sea surface temperature value: [289.98] kelvin

Level-4 data: NOAA 1/4° Daily Optimum Interpolation Sea Surface Temperature (OISST)#

Map of global sea surface temperatures

The NOAA 1/4° Daily Optimum Interpolation Sea Surface Temperature (OISST) incorporates observations from different platforms (satellites, ships, buoys and Argo floats) into a regular global grid. The dataset is interpolated to fill gaps on the grid and create a spatially complete map of sea surface temperature. Satellite and ship observations are referenced to buoys to compensate for platform differences and sensor biases.

Data characteristics#

Spatial resolution: 0.25° latitude × 0.25° longitude.
Temporal coverage: Daily values from September 1981 to the present.
Updated: Daily

Note: Data less than 15 days old may be subject to revision by NCEI

Useful Links#

Dataset Information Page
https://www.ncei.noaa.gov/products/optimum-interpolation-sst
Dataset Landing Page
https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.ncdc:C00844#:~:text=The NOAA 1%2F4 degree,gaps filled in by interpolation.
THREDDS Catalogue
- Human interface: https://www.ncei.noaa.gov/thredds/catalog/OisstBase/NetCDF/V2.1/AVHRR/catalog.html
- Machine interface: https://www.ncei.noaa.gov/thredds/catalog/OisstBase/NetCDF/V2.1/AVHRR/catalog.xml

Crediting the Data Providers#

When using this dataset in publications or presentations, please provide the following citation:

Richard W. Reynolds, Viva F. Banzon, and NOAA CDR Program (2008): NOAA Optimum Interpolation 1/4 Degree Daily Sea Surface Temperature (OISST) Analysis, Version 2. [indicate subset used]. NOAA National Centers for Environmental Information. doi:10.7289/V5SQ8XB5 [access date].

Writing the data to a CSV file#

You can write all or a subset of the data to a pandas dataframe that you can export as a CSV file.

desired_date = '2024-02-01'
year = desired_date.split('-')[0]

url = f'https://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/sst.day.mean.{year}.nc'
xrds = xr.open_dataset(url)

ds_one_day = xrds.sel(time = desired_date)
df = ds_one_day['sst'].to_dataframe()
df.head()

		time	sst
lat	lon
-89.875	0.125	2024-02-01	NaN
	0.375	2024-02-01	NaN
	0.625	2024-02-01	NaN
	0.875	2024-02-01	NaN
	1.125	2024-02-01	NaN

df.to_csv('global_sea_surface_temperatures.csv')

Plotting the data for a day of our choosing#

Let’s use the code we used above to first isolate the data for a single day, then let’s build on that to quickly plot the data.

desired_date = '2024-02-01'
year = desired_date.split('-')[0]

url = f'https://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/sst.day.mean.{year}.nc'
xrds = xr.open_dataset(url)

ds_one_day = xrds.sel(time = desired_date)
ds_one_day['sst'].plot()
plt.show()

_images/7ead9867ca9ccce5cf20a8fd36865b848b37e92d7b330d4c0012cada08c0739c.png

If you accidentally select a date that isn’t present in the time series, you can use method = 'nearest' to select the data from the nearest timestamp to the date you provide. Other methods can also be used, you can read about this here:

https://docs.xarray.dev/en/latest/generated/xarray.Dataset.sel.html#xarray-dataset-sel

ds_one_day = xrds.sel(time = desired_date , method='nearest')

Without the coastlines these data are difficult to interpret. Let’s now explore how to improve this plot and perhaps use some different projections.

Full working example#

Below is a full working example you can copy and play with.

Map projections: https://scitools.org.uk/cartopy/docs/v0.15/crs/projections.html
Matplotlib built-in colourmaps: https://matplotlib.org/stable/users/explain/colors/colormaps.html
Colourmaps for oceanography from cmocean: https://matplotlib.org/cmocean/

desired_date = '2024-02-01'
year = desired_date.split('-')[0]

projection = ccrs.Mollweide()
transform = ccrs.PlateCarree()

fig = plt.figure(figsize=(16, 8))
ax = plt.axes(projection=projection)

# Load and plot the sea surface temperature data
url = f'https://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/sst.day.mean.{year}.nc'
xrds = xr.open_dataset(url)

xrds = xrds.sel(time = desired_date)
sst = xrds['sst'].where(xrds['sst'] > -3)

vmin = -1
vmax = 34

sst.plot(ax=ax, transform=transform, vmin=vmin, vmax=vmax, cmap='jet')

ax.coastlines()
ax.add_feature(cfeature.LAND, facecolor='grey')

plt.savefig(f'sea_water_temperatures.png', transparent=True)
plt.show()

_images/389c0f3cd9a0f8924ce36d034cc7a64acdbc4700968484a4049507487970b995.png

Adding a mask for sea ice#

The dataset also includes sea ice concentration data. Let’s mask the sea surface temperature data anywhere that the sea ice concentration is greater than 0.

desired_date = '2024-02-01'
year = desired_date.split('-')[0]

projection = ccrs.Mollweide()
transform = ccrs.PlateCarree()

fig = plt.figure(figsize=(16, 8))
ax = plt.axes(projection=projection)

# Load and plot the sea surface temperature data
url = f'https://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/sst.day.mean.{year}.nc'
xrds = xr.open_dataset(url)

xrds = xrds.sel(time = desired_date)
sst = xrds['sst'].where(xrds['sst'] > -3)

vmin = -1
vmax = 34

sst.plot(ax=ax, transform=transform, vmin=vmin, vmax=vmax, cmap='jet')

# Load and plot the sea ice
url = f'http://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/icec.day.mean.{year}.nc'
xrds = xr.open_dataset(url)

xrds = xrds.sel(time = desired_date)
ice_mask = xrds['icec'].where(xrds['icec'] > 0)

ice_mask.plot(ax=ax, transform=transform, vmin=1, vmax=2, cmap='Greys', add_colorbar=False)

ax.coastlines()
ax.add_feature(cfeature.LAND, facecolor='grey')

plt.savefig(f'sea_water_temperatures.png', transparent=True)
plt.show()

_images/8ade242a567043012c8f1d48cd2990e8873088eeea1220e99c2b71bdfe22e8e8.png

Finding the sea surface temperature at a given location and date#

Let’s now look at how to find the value for the sea surface temperature at a given location and time. We will using the method nearest to find the nearest value to our selected point, though other methods are available.

https://docs.xarray.dev/en/latest/generated/xarray.Dataset.sel.html

desired_date = '2024-02-01'
desired_latitude = 73.5624
desired_longitude = 30.0142

year = desired_date.split('-')[0]

# Load and plot the sea surface temperature data
url = f'https://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/sst.day.mean.{year}.nc'
xrds = xr.open_dataset(url)

sst_value = xrds.sel(
    time = desired_date,
    lat = desired_latitude,
    lon = desired_longitude,
    method = 'nearest'
)['sst'].values

print(sst_value)

3.87

We might also want to find out what latitude, longitude and time that value corresponds to.

desired_date = '2024-02-01'
desired_latitude = 73.5624
desired_longitude = 30.0142

year = desired_date.split('-')[0]

# Load and plot the sea surface temperature data
url = f'https://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/sst.day.mean.{year}.nc'
xrds = xr.open_dataset(url)

single_point = xrds.sel(
    time = desired_date,
    lat = desired_latitude,
    lon = desired_longitude,
    method = 'nearest'
)

latitude_value = single_point['lat'].values
longitude_value = single_point['lon'].values
time_value = single_point['time'].values
sst_value = single_point['sst'].values

print('Latitude:',latitude_value)
print('Longitude:',longitude_value)
print('Time:',time_value)
print('Sea surface temperature value:',sst_value)

Latitude: 73.625
Longitude: 30.125
Time: 2024-02-01T00:00:00.000000000
Sea surface temperature value: 3.87

Interactive plot#

We can create an interactive plot that we zoom in on, rotate, and observe exact values for the sst at any coordinate. We will use plotly to do this.

Interactive plots can require a lot of memory, so let’s downsample the data first to make it easier to work with for this example.

# Define parameters
desired_date = '2024-02-01'
year = desired_date.split('-')[0]

# Load the dataset from the remote server and select the desired date
url = f'https://psl.noaa.gov/thredds/dodsC/Datasets/noaa.oisst.v2.highres/sst.day.mean.{year}.nc'
xrds = xr.open_dataset(url)
xrds = xrds.sel(time=desired_date)

# Downsample the dataset by selecting every 4th point along lat and lon
xrds_ds = xrds.isel(lat=slice(None, None, 4), lon=slice(None, None, 4))

# Apply a filter to the sea surface temperature data
sst = xrds_ds['sst'].where(xrds_ds['sst'] > -3)

Now let’s extract our latitude, longitude and sea surface temperature values.

We first need to create 2D grids for our 1D latitude and longitude arrays, to match the sea surface temperature values.

sst_array = sst.values  # shape: (n_lat, n_lon)
lon_values = sst['lon'].values  # 1D array of longitudes
lat_values = sst['lat'].values  # 1D array of latitudes

lon_grid, lat_grid = np.meshgrid(lon_values, lat_values)

Next, we will flatten all the 2D arrays so we have 1D arrays of equal length.

lon_flat = lon_grid.flatten()
lat_flat = lat_grid.flatten()
sst_flat = sst_array.flatten()

Finally, let’s plot the data. We will create a ScatterGeo plot that allows us to use a projection. We can select projections from:

https://plotly.com/python/map-configuration/

# Create a scattergeo plot
fig = go.Figure(data=go.Scattergeo(
    lon = lon_flat,
    lat = lat_flat,
    mode = 'markers',
    marker = dict(
        color = sst_flat,
        colorscale = 'Jet',
        cmin = np.nanmin(sst_flat),
        cmax = np.nanmax(sst_flat),
        colorbar = dict(title="Temperature (°C)"),
        size = 6,  # Adjust size for best visual appearance
        opacity = 0.7
    ),
    hovertemplate = 'Lat: %{lat}<br>Lon: %{lon}<br>SST: %{marker.color:.2f}°C<extra></extra>'  # Display lat, lon, SST on hover
))

# Update layout with an orthographic projection and some geographical features
fig.update_layout(
    title = "Sea Surface Temperature (°C)",
    geo = dict(
        projection = dict(type = "orthographic"),
        showcoastlines = True,
        coastlinecolor = "black",
        showland = True,
        landcolor = "rgb(229, 229, 229)",
        showcountries = True,
        countrycolor = "black",
    )
)

# For display in a Jupyter Book (uncomment fig.show() if using a notebook)
# fig.show()
fig.write_html("interactive_plot_sst.html")

from IPython.display import IFrame

# Embed the saved interactive_plot.html file in the next cell
IFrame('interactive_plot_sst.html', width=900, height=600)

Global Sea Surface Temperature (SST) data

Contents

Global Sea Surface Temperature (SST) data#

What Do We Mean by Sea Surface Temperature?#

Measuring Sea Surface Temperature from Space#

Infrared Radiometers#

Microwave Radiometers#

Data Processing Levels#

Level-2: Native Resolution SST Products#

Level-3: Gridded SST Products#

Level-4: Gap-Filled and Model-Interpolated SST#

Visual Comparison of Level-3 and Level-4 Data#

Preparing our Python session#

Level-3 data: OSI-201-b#

Data characteristics#

Useful Links#

Crediting the Data Providers#

Opening and understanding the data#

Writing the data to a CSV file#

Converting from Kelvin to Celsius#

Plotting the data for a day of our choosing#

Plotting sea ice#

Zooming in on area of interest#

Finding the sea surface temperature at a given location and date#

Level-4 data: NOAA 1/4° Daily Optimum Interpolation Sea Surface Temperature (OISST)#

Data characteristics#

Useful Links#

Crediting the Data Providers#

Opening and understanding the data#

Writing the data to a CSV file#

Plotting the data for a day of our choosing#

Full working example#

Adding a mask for sea ice#

Zooming in on area of interest#

Finding the sea surface temperature at a given location and date#

Interactive plot#