Climate Change 101: Observing the climate itself

Roughly since the early 1800’s, people have recorded and transcribed weather observations. Since then, new instruments and methods have entered the science, as well as the idea of climate change itself. We went from measuring short-term weather to long-term climate variations.

So how can we justify when the climate is changing? 

As a refresher, remember the difference between weather and climate. To determine a climate, we’re now averaging or evaluating long-term (10 year minimums) weather information to determine if there are any statistically significant (IPCC) variations over time. I discussed this more in another post, here.

Surprisingly (to me at least) there have been national efforts to measure observations for decades. More so, unforeseen challenges have arisen out of this historical data (e.g., we’re using data meant for short-term observations to evaluate long-term change).

Ok, let’s dive in here.

What are climate services?

Climate services are, in other words, information on climate intended to assist those making the decisions (policy makers, stakeholders, landowners, organizations) aka the decision-makers. It’s like an environmental consultation.

These services need to engage their users and remain accessible. High-quality data from national and international service providers include databases on temperature, rainfall, wind, soil moisture, ocean conditions, maps, risk and vulnerability analyses, assessments, and long-term projections. Depending on the decision-maker, some databases will include complementary information such as demographics, health trends, and other socio-economic variables.

How are we collecting these measurements?

Observations used in weather and climate services come from all kinds of devices. We have instruments ranging between Earth’s surface and atmosphere including buoys and ships, aircraft carriers, and satellites.

Two main categories exist for gathering climate data: in situ, which gathers data at the point of measurement (e.g., sticking a thermometer in the dirt for soil temperature), and remote sensing, which remotely gathers data (e.g., sensors on satellites that infer temperature through radiation refections). In some datasets, in situ and remote sensing is merged.


Dilemma 1: Most atmospheric sensors were designed for short-term weather monitoring and prediction, not long-term climate applications.

Most organizations providing climate services were established before society recognized climate variations and established climatic science, which was approximately in the 1970’s. How we first recognized and developed the science of climate change is reviewed in another post, here.

Dilemma 2: Instruments have been improved, replaced, and have not gathered continuous data when being swapped out.

Continuous data is important when evaluating long-term variations. However, since the original purpose of these instruments was to monitor short-term weather, it wasn’t a priority to avoid introducing “gaps in data”, which is when the instrument stopped for repair or replacement and time would pass with no data recorded.

Dilemma 3: Instruments have been improved with remarkable accuracy, which impairs how we can compare modern measurements to historical ones.

Inconsistency is a huge factor in data reliability. On top of instruments being removed for repair with no temporary replacement, we are now using modern equipment and comparing it to older ones.

Understand how data is recorded, what limitations exist, and you’ll trust the results.

Consider this: How to record climate data.

What should you take into account when planning research? What considerations should you take? The list is long…

  1. What is the spatial scale of the climate phenomena being studied? This tells you how big of a network you’ll need. Are you on the micro, local, meso, synoptic or planetary scale?
  2. What is the temporal scale? If you want to record a finer-scale phenomena, which typically have shorter duration, you’ll need to have equipment that records at a high resolution.
  3. What best expresses your phenomenon? If its humidity, consider if you want to record relative, specific, absolute humidity or dewpoint pressure.
  4. Is the expression (parameter) vector or scalar? Vectors have both magnitude and direction (e.g., wind), while scalar only has magnitude (e.g., temperature).
  5. What instruments will you use and how stable or sensitive are they?
  6. What is the instrument footprint? Or in other words, what can the instrument “see”?
  7. What is the instrument’s response time? If the phenomena is short but the records are taken in long intervals, you may miss key details during the event.
  8. How will the instrument be calibrated, checked for errors, or degrade (drift) overtime?
  9. What does the geography surrounding your equipment look like? You’ll want a geographic representation that is homogeneous and unlikely to have outside influence or change over time, such as urban construction.
  10. What can you include in your metadata, which describes the data you’re recording. This includes the exposure, type and condition of equipment, equipment height, and other protocols.

Ok, but climate variations happen over periods longer than 10 years, right?! How can you record the past?

Archived historical observations are life

You can’t time travel. You need to find trends that include data from before you were born. What do you do? Go.

Climatology and climate science heavily rely on historical records. These include spatial variations in climate over large areas. But again, be aware of limitations.

Limitation 1: Spatial and temporal resolution of historical data may not be fine or accurate enough to answer your question.

This is caused by uneveness in the Earth’s surface, perhaps, which earlier tech couldn’t handle. There are even fewer observations over tropical or polar regions, which may make it hard to make comparisons, and coverage over time is uneven as well.

Limitation 2: Recordings were made by various instruments over time.

Again, the problem with inconsistency in the types of observation systems can alter the reliability of your data.

Limitation n=#! Changes in instrumentation, station location, exposure, practices in observation, and time of observations all combine into quite a distasteful mess.

Take home message

Let’s face it, no one can make that chocolate cake quite like your grandma did… but it still tastes good! These considerations shouldn’t completely destroy your trust in data collection, but instead should create a healthy exercise of practicing caution. A lot of effort goes into gathering, cleaning, and providing data — not to mention the support for those using it!

Next, let’s cover the networks and observation stations in the United States that are providing these climate services.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s