Measurement primer
What a household measurement programme actually measures: the instrument classes, their units, their frequencies and their honest accuracies, plus how to state an uncertainty.
Read this before you write Assignment 1. It is short on purpose: what follows is the vocabulary the assignment expects, not a catalogue of products. Choosing instruments is your job; knowing what a register row has to carry is this page's.
The rule the whole programme runs on
Every measurement exists to settle a decision. Write the decision first — "connect the house roof or leave it" — then the measurement that would settle it, then the method, place, units, frequency and accuracy. A row you cannot attach to a decision is a row to delete, and the money it costs is money you can spend on one that matters.
The programme has a budget of its own: $1,500 for equipment and services across the whole twelve months, separate from the implementation allowance. Your hours are scheduled, not charged.
Water: rain, tank and use
Decides: storage volume, which roof zones to connect, and whether a pump is needed at all.
| What | Method | Units | Frequency | Accuracy, honestly |
|---|---|---|---|---|
| Rain at the site | Tipping-bucket gauge, or a manual gauge read daily | mm | Continuous, totalled monthly | ±2 % to ±5 % of total; under-reads in high-intensity rain |
| Tank level | Pressure or ultrasonic level sensor; manual dip as the check | mm of depth, converted to kL | Hourly logged, or weekly by dip | ±1 % to ±3 % of full scale — on a 5 kL tank that is ±50 to ±150 L |
| Household use | Pulse meter on the incoming main; sub-meters on hot and on the garden line | L, totalled daily | Continuous | ±2 % at normal flows; poor below the meter's starting flow |
| Peak flow | High-resolution logging on the main, or timed bucket tests at each fixture | L/min | A fortnight of logging, or one afternoon of bucket tests | Bucket test ±5 %; it is the simultaneity you are really measuring |
| Water quality | Laboratory sample, or a field kit for turbidity and pH | NTU, MPN/100 mL, mg/L, pH units | Three or four samples across the year, one after heavy rain | Laboratory reports its own limits; field kits are a category, not a number |
Soil, shade and the growing area
Decides: how much ground is worth planting, and which zones are worth improving.
| What | Method | Units | Frequency | Accuracy, honestly |
|---|---|---|---|---|
| Soil texture and depth | Hand texture in the field; auger or spade pit to rooting depth | texture class, cm | Once, at three or more points per zone | Hand texture is a class, not a number; say who did it |
| pH and electrical conductivity | Laboratory on a composite sample, or a calibrated field meter | pH units, dS/m | Once per zone, repeated if you amend | Field meter ±0.2 pH when calibrated that day |
| Organic carbon and nutrients | Laboratory: Colwell P, nitrate N, organic carbon | mg/kg, % | Once per zone at the start | Laboratory method must be named; the numbers are method-specific |
| Soil moisture | Capacitance probe at two depths per zone | % volumetric | Hourly logged through at least one dry spell | ±3 % volumetric uncalibrated; ±1 % calibrated to your own soil |
| Shade | Hourly photographs from a fixed point, or a sun-path app checked against one real day | hours of direct sun | One clear day per season, minimum | The winter figure is the one that decides anything |
Food and waste logs
Decides: what the insect and compost units could actually be fed, and what the garden produced.
| What | Method | Units | Frequency | Accuracy, honestly |
|---|---|---|---|---|
| Food waste | Kitchen scale and a tallied bin, separated into plant, meat and dairy | kg/week | One full week per season, four weeks total | Household compliance is the error term, not the scale |
| Garden waste and prunings | Weighed at the bin or estimated by volume with a stated bulk density | kg/week | Logged when it happens, seasonally | Volume-to-mass conversion is an assumption — state it |
| Harvest | Kitchen scale, logged by crop | kg | Every harvest, all year | The gaps are the error: a harvest eaten in the garden is never weighed |
| Energy for the existing systems | Plug-in energy meter on each appliance; tariff from the bill | kWh, $/kWh | Two weeks per appliance, one summer one winter | ±2 % on the meter; duty cycle is the uncertain part |
Stating an uncertainty
Three things get confused, and the register should keep them apart.
- Resolution
- The smallest step the instrument reports. A tank sensor reading to the nearest centimetre has a resolution of about 30 L on a slim 5 kL tank, whatever its accuracy claim.
- Accuracy
- How far the reading can be from the truth. Usually quoted as ± a percentage of reading or of full scale — and those two are very different at the bottom of the range, which is exactly where an emptying tank lives.
- The error that actually dominates
- Almost never the instrument. It is the household forgetting to weigh the scraps, the gauge sited under the eave, the probe in the one spot that drains differently. Say which error dominates your row, and the marker will believe the rest of it.
Carry uncertainty as a range, not a decoration. "Roof yield 18 to 24 kL" supports a decision; "18 kL ± some" does not. Where a range straddles the threshold a decision turns on, that is the finding: the measurement has told you the decision cannot be made yet, and what it would take to make it.
Calibration, drift and gaps
- Check against something independent. A logged tank level against a manual dip; a field pH meter against a buffer solution on the day; a rain gauge against the nearest Bureau station's monthly total, allowing for the distance between them.
- Expect drift and schedule the check. Monthly for anything with a membrane or an optical window, quarterly for the rest, and after every power loss.
- Mark gaps as gaps. Interpolating across a fortnight of missing tank levels is defensible; recording it as a measurement is not. Your client's supplied findings contain exactly this situation, flagged — go and look at how it is written up.