From mid-May through mid-August 2000, a field project called the Cloud Electrification Studies using Aircraft and Radars (CESAR) is to be conducted in the Kansas-Colorado-Nebraska border area in order to observe the frequent severe convective storms that commonly occur in this region.

This page contains both gridded and non-gridded analysis of positive cloud-to-ground (+CG) lightning as observed by the National Lightning Detection Network (NLDN).  Information contained on this page provides a climatological comparison of significant cloud to ground lightning (CG) events in northeastern Colorado and western Kansas.  


For the purpose of this discussion a significant event was defined as one that at some point in its lifetime had a positive CG fraction greater than 50% while at the same time contained a positive flash density greater than, or equal to, 0.04 flashes/km2/hr.

In an effort to compare the number of significant events (defined above) within range of the CSU-CHILL and Goodland, KS radars, a 200km X 200km area centered at each location, was defined. Archived NLDN data from May - Aug of 1996 and 1997, and May - Jul 1998 (Aug data not yet available) were then gridded for each of these areas.

At each location, a 25km X 25km spatial resolution was used, dividing each area into 64 grid boxes. The gridding scheme used a temporal resolution of 30 minutes, yielding a total of 3,072 grid boxes per day, and 1,038,336 grid boxes for the entire 11 month period studied.

Each grid box was then checked to see if it met the significant storm criteria described above. Once these significant grid boxes were identified, animations of the NLDN data within each area were then generated, using the same 30 minute temporal resolution. Manual analysis was then used to identify the number of significant events (see graphs) associated with several significant grid boxes which may or may not have been spatially or temporally associated with one another other.  

Additionally, of the significant events identified, we have also identified the number of grid boxes and events that contain a grid box averaged +PKI > 75kA (LPC+CG events).  It is important to note that the CSU-CHILL had no grid boxes with average +PKI > 75kA, in fact only 10 grid boxes (or 8 actual events) had an average +PKI > 50kA. The peak average +PKI was 71.35kA.  At GLD, however, we found that 19 of the grid boxes identified as "significant grid boxes" contained average +PKI > 75kA, this translates into a total of 8 actual events that could be defined as LPC+CG events.

The table below provides a brief summary of the data obtained in this analysis (note the data is for the entire observational period).

Total # of Days with

1 or more Sig. Events

Total # of Sig.

Grid Boxes

Total # of

Sig. Events

Total # of LPC+CG













In 1995 the NLDN underwent a network-wide upgrade in order to (1) increase the location accuracy of CG flashes and (2) increase the detection efficiency of low peak current CG flashes down to 5kA (Cummins et al. 1998). Results from Zajac and Rutledge, 1998 indicate that the upgrade did succeed in increasing the detection of low peak current CG flashes; both mean positive and negative peak currents showed decreased values from 1994 (pre-upgrade) to 1995 (post-upgrade). While the findings listed above were expected from the upgrade, the detection of a large population of low peak current positive CG flashes over localized areas was not expected. It has been suggested that low peak current positive CG flashes on the order of a few kiloampers are not real, but rather false detections of IC lightning (e.g., Lopez et al. 1991). Some researchers have proposed using a 7kA peak current threshold for positive CG lightning in order to eliminate the false detection of IC lightning (S. Rutledge, personal communications, 1998).

For the entire area of interest, during the 11 summer months analyzed, there were a total of 1,277,056 cloud to ground lightning flashes, 12.1% of which had positive polarity (above the 7kA threshold). Of those 154,489 positive flashes, 7% (11,923) had peak currents greater than or equal to 75kA.

Looking on a more comparative level between two specific radar locations (CSU-CHILL and Goodland, KS) we find nearly a factor of 3 difference in the amount of positive CG lightning, and over a factor of two difference in the amount of +CG lightning with peak currents greater than or equal to 75kA. For the same time period, CSU-CHILL had a total of 165,375 cloud to ground lightning strikes compared to 118,098 at Goodland. Of these CG strikes, 7.99% of them were positive CGs at CSU-CHILL compared to 21.25% at Goodland. Similarly, we see the number of positive CGs with peak currents greater than or equal to 75kA are significantly less at CSU-CHILL when compared to those at Goodland; 593 (0.36% of all CGs) versus 2,572 (2.18% of all CGs), respectively.


Cummins, K.L., M.J. Murphy, E.A. Bardo, W.L. Hiscox, R.B. Pyle, and A.E. Pifer, 1998: A combined TOA/MDF             Technology Upgrade of the U.S. National Lightning Detection Network.  J. Geophys. Res., 103, 9,035-9,044.

Lopez, R.E., M.W. Maier, and R.L. Holle, 1991: Comparison of the signal strength of positive and negative            c             cloud-to-ground lightning flashes in northeastern Colorado. J. Geophys. Res., 96, 22,307-33,318.

Stolzenburg, M., 1994: Observations of High Ground Flash Densities of Positive Lightning in Summertime                              Thunderstorms.  Mon. Wea. Rev., 122, 1,740-1,750.

Zajac, B.A., S.A. Rutledge, 1998: Climatological characteristics of cloud-to-ground lightning activity in the contiguous            United States, MS Thesis, Dep. Atmos. Sci., Colorado, State, Univ. 119pp.