Calibrating a School Weather Station
A campus weather station reports temperature, relative humidity, and pressure every minute. Over one week in late spring, the station records a daily temperature cycle peaking near 15:00 and a slow pressure fall of about 6 hPa over three days, followed by a wind shift from south to northwest and a 9 °C temperature drop within six hours. A student notices that the humidity sensor reads 100% on several clear afternoons and that the temperature sensor sits inside a white louvered box mounted 1.2 m above a dark asphalt lot. Another sensor in a grassy field 200 m away reads 2-3 °C cooler at midday. The class must decide which records can be used for a climate comparison and how to interpret the pressure and wind event.
1. Diagnose the humidity readings and the temperature discrepancy. Which instrument problems are likely, and what corrections or exclusions are justified?
Read the worked answer
Relative humidity near 100% on clear afternoons is physically implausible unless the air is near saturation; with daytime heating and no cloud, relative humidity normally falls. This points to sensor drift, condensation on the element, or a calibration offset at the high end. The reading should be flagged, checked against a psychrometer or a nearby reference, and excluded from analysis until recalibrated.
The 2-3 °C midday gap between the asphalt-lot box and the grassy field is a siting effect, not instrument error. Asphalt stores and reradiates heat, and a louvered shelter reduces but does not eliminate radiative and conductive bias. Standard practice places temperature sensors over open ground, well away from pavement and buildings. The asphalt-site record is therefore unsuitable for climate comparison but may be useful for local microclimate description. Corrections should not be invented by subtracting a fixed offset; the defensible action is to document the bias, use the grassy-field record for climate purposes, and treat the lot record as a separate urban-heat observation.
2. Interpret the three-day pressure fall, the wind shift, and the rapid temperature drop as a coherent weather sequence.
Read the worked answer
A steady 6 hPa fall over three days indicates an approaching low-pressure system. As the low nears, southerly winds ahead of it bring warmer, moister air; the wind shift to northwest marks passage of the associated cold front. The 9 °C drop in six hours is consistent with cold advection behind the front, and pressure typically stops falling and begins rising after frontal passage.
The sequence is therefore not three unrelated observations but one event: pre-frontal southerly flow and falling pressure, frontal passage with wind veer, then post-frontal cold advection. A prediction that follows is clearing skies, rising pressure, and cooler, drier air behind the front. The class should present the data as a time series so the ordering is visible, since the causal story depends on which change came first. Mistaking the pressure fall for instrument drift would be an interpretive error; the accompanying wind and temperature changes confirm a real synoptic event.
3. The class wants to compare this spring with the same week ten years ago using only this station. What limits that comparison, and what would strengthen it?
Read the worked answer
The comparison is weak for several reasons. The station is new, so there is no ten-year record from the same instrument; sensor drift, shelter changes, and site changes cannot be ruled out. A single week is dominated by synoptic variability, so one spring week says little about climate. The asphalt siting biases absolute temperatures, and the humidity sensor is unreliable, so any derived quantity such as average humidity or heat index is suspect.
Strengthening options: compare against a long-record official station nearby with documented siting and calibration history; use anomalies (departures from each station's own baseline) rather than raw values, so siting offsets partly cancel; and aggregate over a full season or multiple years rather than one week. The class should frame the result as a local case study, not a climate trend, and state the siting and record-length limitations explicitly. Claiming a warming or cooling trend from one week at one uncalibrated site would overreach the evidence.