US Climate Division Data: Caveats
- HDD and CDD often will have "0" mean for certain regions and/or seasons. Ranks and percentiles may yield odd results if all or most of the values are 0. Please keep that in mind in interpreting those plots.
- Recent data is subject to correction at NCDC. Generally, this occurs within 6 months if at all.
- Old Version August 1927 has incorrect temperature values for all of Indiana. These values are included in the climatologies that extend over 1927 and of course in any plots that include that month.
- Old Version Data before 1931 was calculated with a different method. Divisional data was calculated "from linear regression ofstatewide averages to direct climate division averages beginning in 1931". See the A New Homogenized Climate Division Precipitation Dataset for Analysis of Climate Variability and Climate Change: D.B. McRoberts and J.W. Nielson-Gammon, June 2011, Journal of Applied Meteorology and climatology, 50, 1187-1199 and Guttman and Quayle, 1996:A historical perspective of U.S. climate divisons. Bull Amer. Meteor. Soc. 77, 293-303.