Most pages below have links to code and instructions to recreate the results.
|
Hurricanes Notable Trends in Atlantic hurricanes |
|
Hurricane Detection Trends and Corrected Records. The percentage of hurricanes that move near land should be flat. Use that fact to adjust the counts of past hurricanes and ACE for the year to account for missing detection. |
|
Hurricane Category Percentages The percentage of hurricanes in various categories should be flat except for natural variations, global warming effects, and changes in detection methods. |
|
Temperature Trends. No splicing, no homogenization. Just use long records stations and calculate the trends of the highest monthly temperatures: full set of plots and station list. There are increasing maximum temperatures in many locations out west except the immediate coast (within a few miles). Otherwise no trends east of the Rockies. Some downward trends in maximum temperatures around the lakes e.g. Ohio. The Gulf of Mexico and Great Lakes may be moderating extreme high temperatures because of more moisture. Temperature Plots. There are ways to avoid the need for homogenization (discontinuity detection). Use extreme highs, i.e. don't calculate averages or count the number of days since that can be biased by observation time discontinuities (code: monthly-extremes.py) These are plots of the highest summer month temperatures with trends in degrees F per year. All plots on one page including the increases in winter month low temps. |
| Kentfield CA, Tooele UT, Bozeman MT, Lamar MO Farmington ME Charleston SC Lake City FL |
| Dry Season. Similar to temperature I try to use the longest continuous stations for rainall plots. The data shows a general increase in the length of the dry season although in some locations the dry season is shorter and others it has shifted. |
| length of the dry season |
| Calendar Day Rainfall Trends. Trying to show trends in light to heavy rainfalls is complicated by the fact that rainfall can be timed overnight from one day to the next and show up as two days of lighter rainfall. I double count to catch those, but that introduces some trend artifacts in some cases. I also plot weekly rainfall trends and note the uptick in two-inch weeks around the DC area. |
| Rainfall Trends |
| Hourly Rainfall Trends (1-6 hours). Trends in 1 to 6 hour rainfall using hourly data from coop stations. There is no trend in flash-flood producing short duration rainfalls. |
| Hourly Rainfall Trends |
| Strong Tornado Trends There is a statistically significant downward trend in strong tornadoes overall. Off-season (Nov through Feb) show increases using recent data (1990-2020) but those increases are not statistically significant. Also using older data e.g. 1950-2020 those increases go away. |
| Strong Tornadoes |
| Drought No trend in US drought since 1895. 1930's are still unequaled. Even 1976-77 is unequaled. |
| US drought since 1895 |
| Greenland "tipping point" prediction |
| Global Warming White Paper and Misc |
