Note (verbatim, from paper/manuscript_marine.qmd around line 327, under ## Aggregating across station pairs {#sec:uncertainty}, right before Fig. 3 demo_3_uncertainty.png):
did we create a synthetic network? if so, generate some plots that might represent
a southern California, san gabriel valley example with station pairs of appropriate
distances, typical seasonal signals (and inchoerence noise) to demonstrate
how to do a network average with various dvv time series - use the Clements2018
context as an example. The figures demo_3_uncertainty.png does not show how
different individual pairs might give for a dv/v time series and thus understimate the actual
variation.
Context: the surrounding section (lines 315-325) argues that aggregating many station pairs into a network dv/v series exposes the least-reported choice — how to summarize uncertainty (coherence-weighted SE, unweighted SE, between-pair SD) — and that these conventions can disagree by a factor of sqrt(N). The note flags that the current nine-pair synthetic network figure doesn't show realistic per-pair variability, likely underestimating true spread.
Acceptance criterion: a new or revised synthetic-network figure modeled on a realistic southern California / San Gabriel Valley station-pair geometry (appropriate inter-station distances), with a typical seasonal dv/v signal and per-pair incoherent noise, using the groundwater-monitoring context of Clements2018 as the physical template — showing the actual spread across individual pairs (not just the network-average conventions demo_3_uncertainty.png currently compares). Closes when this figure/panel is added and the text/caption references the realistic per-pair spread it demonstrates.
Note (verbatim, from
paper/manuscript_marine.qmdaround line 327, under## Aggregating across station pairs {#sec:uncertainty}, right before Fig. 3demo_3_uncertainty.png):Context: the surrounding section (lines 315-325) argues that aggregating many station pairs into a network dv/v series exposes the least-reported choice — how to summarize uncertainty (coherence-weighted SE, unweighted SE, between-pair SD) — and that these conventions can disagree by a factor of sqrt(N). The note flags that the current nine-pair synthetic network figure doesn't show realistic per-pair variability, likely underestimating true spread.
Acceptance criterion: a new or revised synthetic-network figure modeled on a realistic southern California / San Gabriel Valley station-pair geometry (appropriate inter-station distances), with a typical seasonal dv/v signal and per-pair incoherent noise, using the groundwater-monitoring context of Clements2018 as the physical template — showing the actual spread across individual pairs (not just the network-average conventions demo_3_uncertainty.png currently compares). Closes when this figure/panel is added and the text/caption references the realistic per-pair spread it demonstrates.