Note (verbatim, from paper/manuscript_marine.qmd around line 414, in # Choices that create spurious \dvv\ {#sec:artifacts}, between the clock-drift/seasonal-noise discussion and ## Causal and acausal branches {#sec:branches}):
we briefly mention method A and B, these refer to estimating \dvv\ at each window pair
by selecting the max cc, then doing the averaging of these dv/v time series or if
the CC(t,epsilon) matrix is stacked/averaged, then the \dvv\ is extracted from the
max of the averaged matrix. these may have an effects on the resulting \dvv.
. But we should
also quantify the effect o
Context: "Approach A" (peak-pick each component's CC(epsilon,t) curve, then average per-component dv/v) and "Approach B" (average the CC(epsilon,t) images across components first, then peak-pick once) are already introduced in §sec:aggregation (lines ~293-301, referencing Fig. fig:aggregation), where the text states they "propagate uncertainty along incompatible pathways" but does not quantify the resulting bias/variance difference. Note: the paragraph containing this note is itself truncated mid-sentence ("But we should also quantify the effect o") and needs to be finished/rewritten regardless of the analysis outcome.
Acceptance criterion: a quantitative comparison (figure or table) of dv/v recovered via Approach A vs. Approach B on the same synthetic dataset, reporting the bias and variance of each relative to ground truth — extending the existing demo_2_aggregation.png analysis. Also requires completing/rewriting the truncated sentence in the manuscript once the analysis is in.
Note (verbatim, from
paper/manuscript_marine.qmdaround line 414, in# Choices that create spurious \dvv\ {#sec:artifacts}, between the clock-drift/seasonal-noise discussion and## Causal and acausal branches {#sec:branches}):Context: "Approach A" (peak-pick each component's CC(epsilon,t) curve, then average per-component dv/v) and "Approach B" (average the CC(epsilon,t) images across components first, then peak-pick once) are already introduced in §sec:aggregation (lines ~293-301, referencing Fig.
fig:aggregation), where the text states they "propagate uncertainty along incompatible pathways" but does not quantify the resulting bias/variance difference. Note: the paragraph containing this note is itself truncated mid-sentence ("But we should also quantify the effect o") and needs to be finished/rewritten regardless of the analysis outcome.Acceptance criterion: a quantitative comparison (figure or table) of dv/v recovered via Approach A vs. Approach B on the same synthetic dataset, reporting the bias and variance of each relative to ground truth — extending the existing
demo_2_aggregation.pnganalysis. Also requires completing/rewriting the truncated sentence in the manuscript once the analysis is in.