|In-house quantitative precipitation estimates (a QPE example from Frances of 2004 is seen on the left), radar imagery, satellite imagery, daily weather maps, and tropical cyclone reports are used to figure out brackets of dates where tropical cyclone rainfall is deemed to have fallen, either within latitude/longitude boxes during the real-time graphics over the past several years, or state by state in older cases. The rainfall data comes from the National Climatic Data Center (NCDC) for many cases, and from local mesonets where there is authorized access to the data, like CoCoRaHS and South Florida water management district. All this data is entered into Excel spreadsheets, and a script is run to process the file for plotting within the nMap software system in use at the Hydrometeorological Prediction Center (HPC). Image Magick software is used to create a .gif file of the rainfall isohyetal map, and Paint software is used to create to color-filled graphics. If mesonets are used in the storm total graphics, it is stated within the cyclone's rainfall page and within the related spreadsheets. In real time since 2004, the Climate Prediction Center (CPC) rainfall collective has been used (locally known as the Katz files), which includes COOP and USGS rainfall reports. Once the low has formed, rainfall becomes included within the graphics, and continues for as long as the low exists. If a well-defined tropical disturbance predates the low formation, the storm total rainfall window is extended backwards to account for the pre- existing disturbance. The same is true for remnant mid-level shortwaves or mesoscale convective vortices (MCVs) associated with a weakening TC. Tracks for "remnant disturbance" or "incipient disturbance" are included on the rainfall maps. This procedure causes occasional differences between rainfall amounts of the associated graphics/files within this project, and the tropical cyclone reports issued by the National Hurricane Center (NHC).|
||If the precipitation area
between a tropical cyclone and a nearby front is merged, all the rainfall is
included within the graphics. If there is separation between the
cyclone and precipitation along a draping frontal zone well to its north or northeast
(also known as PRE rainfall), the distant PRE rainfall activity is not included in the storm total rainfall. Same is true with nearby upper lows and tropical cyclones. This is why the excessive rainfall in New England, and the resultant flooding, indirectly due to Lili (1996) is not included within the graphics for that system, but rainfall up into the Mid-Atlantic states for Hilda (1964) would be. For the same reason, it's why the precipitation pattern related to Hermine (1998) is much more expansive than the small size of the system might indicate, since the precipitation pattern between the nearby upper low and the tropical cyclone showed no separation. As for Fran (1996), this means the first 8" or so of rainfall across the Mid-Atlantic was not included, since there was separation between the precipitation from the upper low and the approaching tropical cyclone until nearly the time of landfall. As long as a surface low from a former tropical cyclone (extratropical or remnant low) remains discernible, rainfall continues to be included. This has forced continuations of tracks within the Atlantic and Pacific hurricane databases (also known as HURDAT).