P 3.1 TROPICAL CYCLONE TORNADO RECORDS FOR THE MODERNIZED NWS ERA Roger Edwards 1 Storm Prediction Center, Norman, OK 1 Corresponding author address: Roger Edwards, Storm Prediction Center, National Weather Center, 120 Boren Blvd #2300, Norman, OK 73072; E-mail: roger.edwards@noaa.gov 1. INTRODUCTION and BACKGROUND Tornadoes from tropical cyclones (hereafter TCs) pose a specialized forecast challenge at time scales ranging from days for outlooks to minutes for warnings (Spratt et al. 1997, Edwards 1998, Schneider and Sharp 2007, Edwards 2008). The fundamental conceptual and physical tenets of midlatitude supercell prediction, in an ingredients- based framework (e.g., Doswell 1987, Johns and Doswell 1992), fully apply to TC supercells; however, systematic differences in the relative magnitudes of moisture, instability, lift and shear in TCs (e.g., McCaul 1991) contribute strongly to that challenge. Further, there is a growing realization that some TC tornadoes are not necessarily supercellular in origin (Edwards et al. 2010, this volume). Several major TC tornado climatologies have been published since the 1960s (e.g., Pearson and Sadowski 1965, Hill et al. 1966, Novlan and Gray 1974, McCaul 1991, Schultz and Cecil 2009). While these studies undoubtedly have provided valuable insights into TC tornado distribution, the tornado data have evolved markedly during that time span, as have the ways the data are collected. Throughout these changes, and even though TC tornado prediction is evolving out of historically dominant empirical and climatological (Weiss 1987) modes, an understanding of tornado frequency and distribution in TCs remains foundationally crucial to their research and prediction. That understanding can be hindered or misled by often subtle or inconsistent influences of underlying secular artifacts on the data, and by changes in those artifacts with time. Nationwide, tornado data have experienced a pronounced increase in reports over the decades, roughly doubling the mean yearly tally since the mid- 1900s. This trend largely is attributed to a combination of technological advances in documentation capabilities (e.g., digital cameras, camcorders and, more recently, cell-phone imaging), increasing population, greater media attention, proliferation of spotters and storm chasers, and intensified National Weather Service (NWS) emphasis on warning and verification efforts in the WSR-88D radar era. More discussion of such secular influences can be found in Doswell and Burgess (1988), Brooks and Doswell (2001), McCarthy (2003) and Doswell (2007). In addition to covering such factors, Verbout et al. (2006, 2007) quite vividly illustrated that the increase in nationwide tornadoes over five decades since 1954 was attributable to the weakest (F0) bin of damage rating (Fig. 1). This is the very class of tornado that is most common in TC records, and most difficult to detect in the damage above that from the concurrent or subsequent passage of similarly destructive, ambient TC winds. As such, it is possible (but not quantifiable) that many TC tornadoes have gone unrecorded even in the modern NWS era, due to their generally ephemeral nature, logistical difficulties of visual confirmation, presence of swaths of sparsely populated near-coastal areas (i.e., marshes, swamps and dense forests), and the presence of damage inducers of potentially equal or greater impact within the TC envelope. Figure 1. Annual tornado counts (triangles) and tornadoes ≥F1 (dots), 1954-2004. Linear regression lines for each series in red and blue, respectively. Adapted from Fig. 2 in Verbout et al. (2007). As specifically applied to TCs, increasing trends in tornado records were apparent as long ago as the 1960s (Hill et al. 1966). Since TC tornadoes tend to be distributed more toward the low end of the F/EF scales (hereafter, EF) than the national figures at large (Schultz and Cecil 2009, and as demonstrated herein), any enhanced emphasis toward gathering of records of weak (EF0-EF1) tornadoes hypothetically should contribute to increased TC tornado numbers. 2. THE “TCTOR” DATASET a. Justification and tornado-record characteristics Given the aforementioned concerns with the tornado data overall, and by extension, with TC tornado records, a more focused, updated and