Observations are direct measurements of the Earth’s climate system, collected from instruments like weather stations, satellites, buoys, and aircraft, or from natural records (tree rings, ice cores, and sediments). They provide the foundational data for tracking temperature, precipitation, sea level, and many other variables over time. Observational data is essential for understanding current conditions and detecting trends or extremes. Observations are the most direct and accurate representation of what is happening but can have gaps, especially in remote areas, hard to measure variables or over long time scales. Also used for validating models and understanding current trends.

Note that all information given here is collected through individual contributions from ESMO SSG members and other participants. If you would like to suggest some additions or contribute to the page, please do so using this googledoc. https://docs.google.com/document/d/1Tu-7f4RrWCQzGuJ7lTxqb6aiZBJb1V5a0duEsKAp72c/edit?usp=sharing.   If you are not able to access the googledoc, you can also send your suggestions over email.

Important Data links and catalogues

You can find important observational data records through the following links.

Other resources

The Earth Observation Handbook

https://eohandbook.com/eohb-landing/img/CEOS_EO_Handbook_2023_kv_compact_v2.png

The CEOS Earth Observation Handbook presents the main capabilities of satellite Earth observations, their applications and a systematic overview of present and planned CEOS agency Earth observation satellite missions and their instruments.

NASA Eyes on the Earth

Fly along with NASA's Earth science missions in real-time, monitor Earth's vital signs like Carbon Dioxide, Ozone and Sea Level, and see satellite imagery of the latest major weather events, all in an immersive, 3D environment.

CSA page on Earth Observation Satellites

Learn how Earth observation satellites provide data on oceans, ice, land environments, and the atmosphere, or explore different satellite missions that are operated or supported by the Canadian space agency.

Important Observation related terms

These are some concepts that are specifically important in relation to observations.

  • Metadata - Descriptive info about data (e.g., instrument type, location, calibration history) that's crucial for proper interpretation.
  • Proxy Data - Indirect climate evidence (e.g., tree rings, ice cores) used to infer past climate conditions before modern instruments.
  • Quality Control (QC) Flags- Indicators attached to data points that flag potential issues like errors, gaps, or suspect readings.
  • Coverage Bias - Gaps or uneven distribution in spatial or temporal data that affect the representativeness of observations.
  • Reprocessing - Re-analysis of raw observational data using updated algorithms or corrections to improve accuracy and consistency.
  • Ground Truthing - Using direct, on-the-ground measurements to validate or calibrate remote sensing data, often from satellites.
  • Instrument Drift - Gradual changes in instrument response over time, affecting long-term measurement consistency.
  • Homogenization - Techniques used to adjust observational time series to account for non-climate-related changes (like station relocations or equipment updates).
  • Radiometric Calibration - The process of converting raw satellite sensor data into physical units like temperature or radiation.
  • Inhomogeneities - Discontinuities or biases in a climate record caused by changes in instruments, observing practices, or environment.

 

Visualisers and APIs 

Further information and training links

Overview of observation types

Determining what observation you need:

Other constraints to consider

  • To make a useful dataset from measurements, data needs to be proceseed e.g. be on a regular grid, be validated etc.
  • Time period: EO (Earth Obs) only goes back to the 70s at best, and new instruments are much better than older ones. For long records especially, be aware that several satellites must be stitched together (and for even longer records, combined with in-situ obs). This has an impact on which features can be resolved and things like gaps and the variability characteristics of the data through time
  • Timestep: depends on the underlying data source, although the relationship is not necessarily simple, particularly for in situ/point measurements where the representativeness of the observation over a wider area must be considered – an explainer from ARGO here
  • Resolution – data volume as well as your intended purpose is a consideration here.Global/regional  coverage– in addition to your needs, data quality may vary as a function of location (e.g. for many satellite products, observations can only be made for clear sky, so there may be larger uncertainties or fewer observation at some times of year in some locations). For in situ observations, typically the southern hemisphere is less well sampled than the northern 

In situ observations

Data record types 

How we combine the information we collect depends, in part, on what we intend to do with it. (ie. observations are often combined for a specific purpose e.g. data assimilation).

Climate scale observations (WGCM related)

NWP scale observations (WGNE related)

Explainer: Near Real-Time Versus Standard Data Products | NASA Earthdata

Observations for modellers

  • For model evaluation
    • CMIP REF
    • ESMValTool,
    • PCMDI metrics (and the other ref -related packages)
  • Forcing, assimilation and initialisation
    • CMIP forcing pages (coming soon?)
  • High resolution modelling – evaluation
  • Process studies?
  • How are observations used in reanalysis

Other info:

Data Rescue / Citizen Science