Skip to content

Citation

If you use HydroSeason in research, cite the software release.

Software

Cite the version you used. For HydroSeason 0.2.0:

@software{tayer_hydroseason,
  author  = {Tayer, Thiaggo C.},
  title   = {HydroSeason: Remote-sensing-first hydrological year and season detection},
  version = {0.2.0},
  year    = {2026},
  url     = {https://github.com/tayerthiaggo/hydroseason},
  doi     = {10.5281/zenodo.22956109}
}

Version DOI (0.2.0): 10.5281/zenodo.22956109. To cite the project as a whole, across versions, use the concept DOI 10.5281/zenodo.21866898, which the README badge links and which always resolves to the latest release.

Scope note

HydroSeason analyses satellite-derived surface-water extent. Rainfall is ancillary context only and never sets water routing, boundaries, phases, events, or low spells.

Circular-timing precedents and limits

HydroSeason's annual timing statistics are close methodological precedents, not validation that satellite surface-water extent is interchangeable with gauge discharge. Mao et al. provide the satellite surface-water precedent: they describe annual maximum flood-extent timing with circular mean direction and mean resultant length (DOI: 10.1029/2019JD031381).

The following four are flood-discharge or annual-flood studies, not satellite-surface-water validation. Hall and Blöschl use circular concentration and a uniformity assessment for European flood seasonality (DOI: 10.5194/hess-22-3883-2018). Cunderlik, Ouarda, and Bobée show why short annual-maximum records need sampling caution (DOI: 10.1029/2003WR002295). Matti et al. illustrate flood-season shifts across Scandinavia (DOI: 10.1002/hyp.11365), and Villarini documents geographic variation in US flood seasonality (DOI: 10.1016/j.advwatres.2015.11.009).

These papers motivate transparent circular effect sizes and uncertainty reporting; they do not prescribe HydroSeason's package thresholds. The circular Kuiper recurrence conjunction, detectability floors, five-detectable-year guard, seven-cycle annualization guard, and direct-profile refinement thresholds are explicit software decisions frozen under hydroseason-v0.2.0 and are documented in the Methods Reference.

Interpretation limitations

  • Bootstrap intervals resample the usable annual timing observations. They quantify sampling variation under that resampling scheme, not sensor error, spatial error, autocorrelation, or causal hydrology.
  • The Monte Carlo Kuiper p-value tests a discrete 12-month uniform null. It is rotation-invariant and complements R, but does not identify a physical mechanism or make bimodal timing unimodal.
  • Calendar months are the available monthly-resolution timing units. A peak can move within a month and a monthly water mask can miss short floods; results should not be interpreted as daily discharge timing.