Choose the boundary before counting water flows

A process becomes an external flow only at the chosen boundary

The same movement of water can be an output for one system and an internal transfer for a larger one. This investigation includes surface water, soil water and groundwater within one fictional basin store. Its external ledger counts precipitation entering, actual ET leaving to the atmosphere, and total streamflow crossing the outlet.

Draw a large boundary around three smaller boxes labeled surface, soil and groundwater. Add a stream leading to the outlet. Place the precipitation arrow across the outer boundary, then draw infiltration from surface to soil and recharge from soil toward groundwater. Those latter arrows stay inside the selected system.

Water budgets depend on the system and exchanges being studied; surface and subsurface compartments can be accounted for separately or together. USGS Circular 1308 on water storage, exchanges and budgets. The browser uses the combined store deliberately, so it does not subtract an internal transfer as though water vanished from the whole basin.

Follow one parcel without counting it twice

Imagine a water parcel entering as precipitation, infiltrating soil, recharging groundwater, entering a stream as baseflow and eventually passing the outlet. Mark each step on your drawing. It crosses the selected outer boundary on entry and at the outlet. Its intermediate route changes where it is stored, not the total amount within that boundary.

The fixed Q = 40 mm represents total stream outflow for the month, including any groundwater contribution already carried by the stream. Subtracting an additional baseflow amount would double-count that contribution. Subtracting recharge as another external loss would also be inconsistent with including groundwater in the basin store.

This is an accounting rule for the stated system, not a claim that groundwater never crosses a watershed boundary in nature. The model explicitly assumes no such boundary exchange. A real study would need to investigate subsurface connections instead of assuming that a surface drainage divide always defines a closed groundwater system.

Change the box and change the ledger

Now draw a separate box around soil alone. Water moving from soil into deeper groundwater crosses that smaller boundary and can be an outflow from the soil-water budget. The same recharge can be internal to the combined basin budget. There is no contradiction when the boundaries and storage definitions are stated clearly.

For a paper task, color internal arrows one way and boundary-crossing arrows another, then add words so the distinction does not depend on color. Ask another learner to identify the system from your ledger without seeing the drawing. If they cannot, add the boundary, included stores, area and interval to its heading.

Do not add every named water-cycle process as a separate loss merely because it appears in a diagram. Conservation requires tracking the net crossings for the chosen system. An elaborate diagram can still produce a wrong budget if the same water is counted repeatedly.

Closure does not verify the measurements

The default budget has P = 100, ET = 60 and Q = 40 mm, yielding zero storage change. Consider a hypothetical data-processing error that adds 10 mm to both P and ET. The altered values, 110 and 70 mm, still give zero storage change. The balanced result does not reveal those compensating errors.

This counterexample is constructed for reasoning; it is not a claim that the browser’s synthetic values contain measurement errors. It shows why checking a sum and validating each input are different tasks. Independent measurements, uncertainty estimates and consistent definitions are needed to assess a real ledger.

Matching totals also do not establish causation. Two months can share a net change while differing in storm timing, vegetation activity or flow paths. The monthly box model does not recover those histories from one endpoint.

Audit the basis of every number

Write a checklist beside your paper ledger: same basin area, same month, same depth-or-volume convention, explicit storage definition and no duplicated crossings. If one source reports a rate such as cubic meters per second, it must be integrated over the relevant interval before comparing with monthly volumes. The browser already supplies monthly amounts; it does not perform that observational integration.

The practical product is an auditable argument. State what the budget demonstrates, which assumptions make it valid, and which observations would be required before applying it to an actual basin.

Transfer: repair a double-counted ledger

A learner uses the default budget and subtracts an extra 15 mm of internal groundwater-to-stream flow, claiming ΔS = −15 mm. The outlet’s 40 mm total already includes that water. What changes?

Check the watershed-boundary transfer

Remove the duplicate 15 mm subtraction. For the combined basin, ΔS = 100 − 60 − 40 = 0 mm and ending storage remains 100 mm. The groundwater-to-stream transfer can appear in a compartment diagram, but it is not an additional crossing of the selected outer boundary.