(4) With Q3 as reference, CloudSat underestimates (overestimates) certain snowfall when the bin height of detected snowfall events are below (above) 3 km, and generally overestimates light snowfall (<1 mm/h) by 7.53%, and underestimates moderate snowfall (1 2.5 mm/h) by 42.33% and heavy snowfall (⩾2.5 mm/h) by 68.73%.
(5) The bin heights of most (99.41%) CloudSat surface snowfall events are >1 km high above the surface, whereas 76.41% of corresponding Q3 observations are low below 1 km to the near ground surface.
This paper studies a variant of the three-dimensional bin packing problem (3D-BPP), where the bin height can be adjusted to the cartons it packs.
For each robustness bin, we determined the number of genotypes of the sample that fall in that bin; and the jackknife estimate gave the error bar on the height of the bin.
In histograms, the bin width was chosen by data-based kernel smoothing [ 19] and the bin height was scaled to a unit area under the curve [[ 20], p.168].
Then, we generate a histogram of log 10 (F i ), with bins [0, 0.2), [0.2, 0.4), …, [3.8, 4), normalized to total area of 1. Having H a, b denote the height of the bin [ a, b), we define pdf(g) = max 0.0001, H a,b ), where [ a,b) is the bin containing log 10 (g).
