AD//HOC
Assessing Ambient Air Temperature as a Climate Variable
Ambient air temperature, sometimes referred to as average air or room temperature, gives zero insight about the weather forecast; it does not predict or summarize or quantify. It merely communicates what the air temperature is at this moment in relation to a specific location. As such, its value constantly fluctuates minute-by-minute.
Through all that it influences and determines, ambient air is a decisive climactic factor that must be seen as a primary focal point when working with climate and weather data. Air temperature can affect everything from individual comfort to entire industries; crop growth, material longevity, shipping routes (Shariff, 2018), energy use, and even murder rates (Ingraham, 2019) are some of the crucial aspects of our society that can be seen to have a direct connotation with how the air feels. To some it might be seen as simple to measure and basic in concept, but ambient air temperature (mostly referred to hereafter as simply ‘temperature’) can hold an incredible amount of nuances and subjectivity; while collected data might not change, what it means to the observer easily can. From even the most basic education on the topic, it is communicated that weather is a function of both temperature and climate temperature, and it is seen as the baseline for climate and weather analyses. It is something that the world is an observer of-- its extremes can spell doom for entire ecosystems while its subtle shifts and idiosyncrasies can inform decisions around the world.
While accuracy is important to any data collection process, ambient air temperature does not always require the precision that other atmospheric variables might command, such as reflectivity. However, the business problem needs to be kept in mind when considering accuracy; the reason for collecting temperature data can greatly inform how concerned the observer needs to be with accuracy. A data error might be minute in one scenario, whereas the same error can change the outcome entirely under different circumstances. With this in mind, it is still critical to ensure consistent data quality and accuracy through proper verification and quality assurance. Without data provenance, there is no way to confidently present insights, thus rendering the analysis useless.
Temperature is a variable that, due to our sensitivity to it, is easy to comprehend in the abstract. Our bodies might not be able to detect individual degree-changes, but our homeostatic senses are certainly capable of telling us when it's too hot, too cold, getting warmer, or getting cooler. This is why temperature serves as a baseline for all weather forecasts-- from a layman’s perspective, most people view weather as simply a flux between temperatures and precipitation types and amounts. While there are hundreds of ways to underscore the importance of understanding temperature in our day-to-day lives, industry can serve as a good starting point to further analyze temperature and its effects. In Asphalt Materials Science and Technology (Speight, 2016), it is outlined that air temperature is a major index that influences the hot mixed asphalt temperature, and thus is crucial to obtaining ‘compaction’ and longevity of newly paved surfaces and patches. The book goes on further to explain that if the ambient temperature is cooler than required or specified, then “...the asphalt pavement will cool much faster than the best estimates, which causes the mix to set making it very difficult to obtain the required or specified compacted density. The usual requirement for ambient temperature is at least 50°F. If the temperature is below this point, or if there is wind or precipitation, then the asphalt mix will cool and deprive the mix of the desired compaction.” The book continues to outline what temperature-related variables can impact a project, and what to look out for when planning:
1) Awareness and monitoring of the ambient air temperature
2) Awareness and monitoring of the base temperature, and
3) Consistent monitoring of the temperature of the pre-set asphalt
In both the planning and execution of such a task, one must be extremely aware of the ambient air temperature both in real-time and forecasted during the task’s time period. This puts an extreme emphasis on accurate, valid, real-time data, as even something as small as a two-degree discrepancy (49°F measured vs. 51°F actual) could cause the entire undertaking to fail.
Continuing with the asphalt project, there are many methods a firm could utilize to ensure accurate and valid temperature data. Radiosonde and AMDAR observations can give the firm a robust response with regard to above-surface air temperature -- but how can that help us now? Mesonets can provide temperature data at an extremely granular level, but that might be more data than is actually needed. CO-OP stations measure temperature and other variables across the country in real-time, with almost 10,000 CO-OP sites throughout the 50 U.S. states and its territories. However, a firm might be looking for something a bit more official, and while the data readings are known to be accurate for COOP sites, our problem requires a focus on a more local level. In situ temperature data is necessary to get accurate readings for now through the end of the project lifecycle. One reliable source for in situ data is the utilization of Automated Surface Observing Systems (ASOS). ASOS units are operated and controlled cooperatively in the United States by the NWS, FAA, and DOD, and provide a larger spread of data than its sister-stations, AWOS. ASOS systems report at hourly intervals, giving the firm real-time readouts if the risk of a temperature drop is imminent. ASOS stations also report special observations if weather conditions change rapidly and cross aviation operation thresholds. According to the FAA, they generally report parameters related to visibility, precipitation, pressure, and wind-speed (crucial variables for aviation), while also having the capability to report temperature and dew point in degrees Fahrenheit, ice levels, lightning, sea level pressure, and precipitation accumulation. ASOS serves as a primary observing network in the United States, ensuring an extra level of accuracy as many federal agencies are dependent on ASOS readings.
It is safe to operate under the assumption that the firm would receive read-outs from an ASOS as METAR/SPECI data, which includes ambient temperature in degrees Celsius (meaning the danger zone for the asphalt would be 10°C). This data is reported automatically and includes a range for daily temperature and error bounds (Figure 3), and can even be reported in intervals as tight as one minute (due to the One-Minute Observation [OMO] read-out). Human observers seldom interact with the ASOS station (especially any human observer from the firm at hand) unless a back-up or update is needed.
There are 52 AWOS/ASOS stations in Tennessee with 2 in my immediate area (USDOT); one ASOS (at the airport) and one AWOS III. These stations report most of the same units, with the ASOS also reporting reporting temperature and dew point in degrees Fahrenheit, weather conditions, ice, lightning, sea pressure and precipitation accumulation. While there is no inherent problem with the stations, it was strange to notice the number of AWOS stations (44) to the number of ASOS stations (8), especially keeping in mind the different data they collect. Much of Tennessee exists on a floodplain-- the state has had numerous 100-year-floods in the past decade alone-- and it seems odd that the state would have a larger number of AWOS stations over ASOS (Figure 1), while only the latter of which collects any data on precipitation accumulation which is crucial in the reporting of flooding (NOAA).
While its importance as a variable is established, this doesn’t prevent a host of issues from occurring when we try to do something as simple as measure temperature. While it may seem straightforward, there are a number of things to consider when collecting ambient air temperature. When looking at a thermometer, we might recognize that it is 70°F outside if that is indeed the thermometer’s reading. However, this is actually a measurement of the temperature of the sensor itself, not necessarily the ambient air. A thermometer left outside in direct sunlight will result in a reading that is inaccurate and invalid, as the ambient temperature and the heat generated by the absorption of light are measured in tandem. This is why precautions are taken to protect thermometers from outside interference by placing them in protective casings or screens -even ASOS stations are fitted with structures to limit outside influence on temperature reading (see Figure 2).
While it’s always worthwhile to implement quality control procedures on any data source, ASOS read-outs are government data and can be held to a certain standard more-so than crowdsourced weather data. With that, good source quality control procedures would be to accommodate for missing values with supplemental data sources (think RAWS from a nearby area), and use these secondary sources to track any discrepancies over time. Although this wouldn’t necessarily fix the quality of the data, it could help map out any particular changes from the norm that might be worth a second look. This could be implemented with a simple KPI/API set-up; assuming these data sources can both be ingested through an API, as the data comes in, there could be certain flags that go off if one source diverts outside of a set of standard bounds. While this would show comparatively data quality and data symmetry, it can also serve as a measure to indicate if extreme weather, per any of the measured variables, was occurring.
Looking at the big picture, we can deduce that ambient air temperature exists as a dichotomy in the world of atmospheric variables. It affects so many aspects of our lives, yet we are too obtuse to measure incremental changes on our own. We can find our local air temperature in a few clicks, yet it requires often-underutilized precautions for us to ensure it is measured correctly. It is the baseline for weather and climate as we know it, yet it’s situational subjectivity makes it so that every industry and locale is affected differently by it. Even though we all have the same understanding of what 70°F is, it would not be sound analysis to use this variable in all circumstances. To proceed with data collection, I would first have to assess what exactly I’m collecting data for. What specifically do I want to measure, and where do I want to measure it? Over what time? In what increment? While these factors might not affect the temperature as it exists, it changes everything about how we collect and process our data. To continue with the asphalt example, I would utilize ASOS station METAR read-outs via the AWOS/ASOS Data Acquisition System in order to get accurate, real-time readings of the ambient air temperature. While this would provide more data than needed, the data can be trusted to a certain extent while also providing at a consistent rate to keep up with real-time measurements. This might seem like a lot of work for something that is measured in vernacular with simply it’s too hot, but in fact a lot of the nuance and intricacy that goes into capturing ambient air temperature accurately and honestly is dependent on the why. The reason we are collecting the data can inform the data more than we let on.