You're looking at 196,856 water-chemistry samples from one NEON aquatic site, with hundreds of records per site rather than a handful. Pick a tab above, or jump straight to the relationship between two analytes.
How the stats stay honest: every statistic shows its sample size; correlations default to Spearman with an n ≥ 8 reliability flag; the seasonal view is a real STL decomposition of the monthly record, not a forecast; below-detection values are flagged, never hidden; and one per-analyte plausibility gate keeps a single artifact from driving any fit, the trend, the predictor, or the map's colour scale.
Show all analytes
Find sites by their water chemistry across the whole network, then jump to one. Values are each site's average over its full bundled record, computed the same way as the map (implausible extremes excluded). This ranks sites by a within-site index, it is not an absolute ranking, and below-detection samples are kept in the count and flagged, not dropped.