The School-Yard Thermometer and a Four-Week Climate Claim
For a science fair project, Lena wants to show that her city is warmer than the surrounding countryside. She mounts a digital thermometer in a white plastic box on the south wall of her school, 2 m above an asphalt parking lot, and reads it every day at 15:00 for four weeks in July. She compares her readings with the daily maximum temperature from the official weather station at the airport, 12 km away, which logs continuously from a louvered screen 1.5 m above mown grass. Her four-week average of the 15:00 readings is 31.8 C; the airport's four-week average of daily maxima is 28.4 C. The gap is about 4.6 C on clear, calm days and about 1.2 C on windy, overcast days. She writes, 'The city is 3.4 C warmer than the country, and this shows global warming.'
1. Explain the physical reasons her sensor reads higher, and why the gap shrinks on windy, overcast days.
Read the worked answer
Why the sensor reads high. Asphalt has a low albedo and a high heat capacity, so it absorbs a large fraction of incoming solar radiation and then releases it as sensible heat, warming the air immediately above it. The south wall adds two more paths: it conducts heat from the building fabric and radiates infrared onto the box. The box itself, even painted white, is poorly ventilated compared with a louvered screen, so air inside is warmed by the box walls and cannot exchange freely with the outside. At 15:00 the surface is close to its daily temperature peak, maximising all three effects.
Why wind and cloud matter. Wind increases convective exchange, mixing the shallow hot layer above the asphalt with cooler air from elsewhere - the same heat is spread through a much larger volume, so the local excess falls. Cloud reduces incoming solar radiation, so the asphalt absorbs less and the surface-driven heating never develops. Both effects point the same way: the gap is a siting artefact that depends on the weather, which is itself evidence that she is measuring her microsite more than her city.
2. Identify two problems in how she compares the two numbers, and explain how each distorts the 3.4 C figure.
Read the worked answer
Problem 1 - mismatched statistics. Her value is a single instantaneous reading at 15:00; the airport's is a true daily maximum from a continuous logger. On many days the maximum occurs at 13:00 or 16:30, so she is comparing her own spot sample with a value that may come from a different hour entirely. The two numbers are not the same quantity, so their difference has no clean meaning - she should compare daily maximum with daily maximum, or hourly mean with hourly mean, using her own continuous record.
Problem 2 - one month of one season cannot support a climate claim. Four weeks in July samples weather variability, not climate. A single hot July, or a single site, is not evidence about long-term trends, and 'global warming' is a statement about decades of globally averaged data, not a parking-lot thermometer. Her 3.4 C cannot even be called the city's urban heat island intensity, because one urban site and one rural site in one month give no idea how representative either is.
A subtler issue: the airport is 12 km away and may differ in elevation and distance from the coast, so some of the gap has nothing to do with urban or rural character at all.
3. Design a fair comparison and state precisely what claim her data could then support.
Read the worked answer
Instrument and siting. Use identical loggers in identical louvered radiation shields, mounted 1.5 m above ground, over similar surfaces (grass at the reference site, and for the urban site a standard exposure rather than against a wall). Ventilation, height and shielding should match the official standard exactly, because these are what make readings comparable.
Paired, simultaneous operation. Run the urban and rural loggers at the same time, calibrated side by side first to confirm they agree to within 0.2 C in the same place. One urban and two or three rural sites, with several urban sites of different surface type (park, asphalt lot, residential street) would show how much of the effect is surface-driven.
Analysis. Compare daily maxima with daily maxima, daily minima with daily minima, and daily means; report paired differences day by day with a mean and spread, not two averages. Run for at least a full year to capture seasonal variation, and repeat across years before saying anything about trends.
What the data could then support. A statement such as 'over this period, this urban site's daily maximum averaged X C higher than the rural reference, with the largest differences on calm clear days' - a local urban heat island measurement. It still would not support a claim about global warming, which requires long-term, globally distributed records and trend analysis.