Prediction and occurrence of aircraft lightning encounters at Amsterdam-Schiphol Airport.

In this hightlight a study (1) on lightning encounters around Amsterdam Airport is summarized. At Northwest Airlines (NWA) a study (2) was done in 2006 with historic data from NWA aircraft lightning encounters from 1997 through early 2006. The focus of the study was to look at cold season aircraft lightning encounters in relation to the characteristics of the surrounding atmosphere. One result was the discovery that 40% of all lightning encounters with NWA aircraft occurred during the months of October through April, which is not the period of the most frequent thunderstorm activity during the ‘warm season’.

Next, an examination was done of the weather conditions surrounding KLM aircraft during lightning encounters around Amsterdam Airport Schiphol for the ‘cold-season’ months. Located at about 12 km to the east of the North Sea, the climatology of Schiphol shows 13 days with thunderstorms in the cold season. Compared to other commercial airports like Frankfurt (5 days), Brussels and Hamburg (8), Rotterdam (9) and Paris (4) this is remarkably high. KNMI has observed that lightning originating from 10.000 to 15.000 ft cloud-tops (3-5km) is not uncommon and aviation meteorologists have speculated that the aircraft in the vicinity frequently trigger the lightning strikes. KLM-studies (1998–2006) on lightning encounters in the Dutch Airspace show that 84% of all encounters occur in winter at low altitudes around 3000ft and only 16% in summer. For a better understanding an additional study was performed, based on the NWA study2) in the cold season (October 2003 - April 2007). 

In this highlight a summary of this lightning study for Amsterdam Airport is presented first, followed by an introduction of the so called awareness report for Aircraft Induced Lightning (AIL), a new tool to make users aware of the occurrence of this specific weather situation in the Amsterdam Flight Information Region (FIR), and some conclusions.

The Lightning study 
For the study we used archived radiosonde data (twice a day, station De Bilt), METAR data (Schiphol Airport), radar and lightning data (from the KNMI SAFIR/FLITS network) and 62 KLM Pilot Reports. From the De Bilt soundings several instability indices were derived, as well as the 700hPa (Flight Level 100) wind direction and speed and temperature. With the help of the radar data several cloud and precipitation characteristics that are used in aviation were derived and from the SAFIR/FLITS Lightning data the frequency of strikes and correlation with detected precipitation echoes were derived.

As in the Fahey study (2), a strong correlation between stability indices from the nearest sounding observation and the KLM lightning encounters was noted. If only data is considered for wich radar echoes were observed, threshold values for potential convective activity were reached for all stability indices in over 80% of the cases.

In all cases, the temperatures observed at the Lifting Condensation Level (LCL) were above the –100C threshold criteria required for cloud-to-ground strike, but Equilibrium Level (EL) temperatures were generally not cold enough (below -200C) to reach the criteria of the Storm Prediction Center (SPC) of the US National Weather Service (NWS) for the prediction of cloud-to-ground lightning3). However, the soundings in nearly every case where EL and LCL criteria were met, showed a range of temperatures in the clouds between –100C and –200C where ice and supercooled liquid water droplets can co-exist, stimulating the separation of electrical charges during the convective process. It was observed that lightning strikes invariably were recorded from cells with a minimum intensity of 31dBZ (in a range to over 47 dBZ) on the KNMI-radar. In some cases we were able to relate a single recorded lightning strike with the reported aircraft lightning incident.

The convective areas showed little change in intensity and coverage, with only 5% of the cases showing an overall decreasing trend and 25% of the cases showing an increasing trend. About 44% of the cases contained tops of convective clouds at only 15,000 ft or less. 68% of the lightning encounters occurred with a 700hPa flow out of the NW (270 through 360 degrees) quadrant. 26% occurred with a wind from the SW quadrant, mainly between 250-270 degrees. Most of the cases with winds more southwesterly than 250 degrees had thunderstorms with warm-season characteristics. Significant cold-season convection was often triggered over the North Sea, combining cells to form larger cells over land. However some events seemed to be independent of this typical North Sea process. 

Pattern Classification for northwest airflows
A northwest airflow brings unstable and cool or even cold air from higher latitudes over the relative warm North Sea to the Netherlands. Figure 2 is a representation of this typical airflow at 500 hPa. 

The frequency of occurrence of this airflow was computed for the research period 2003 – 2006, cold season only. November to February showed the highest occurrence (29% to 35%), the lowest was in March, April and October (20% or lower). Figure 3 is a satellite imagery of the organized clouds over the North Sea, lined up in cloud streets or open cell structures depending on wind and instability, familiar to the Lake effect over the Great Lakes in the USA. 



Aircraft Induced Lightning (AIL) in the operational Meteorological Service.
To test whether a forecasting tool is a) practical to produce, and b) could be beneficial for avoidance procedures, a prototype forecast for AIL in the form of an experimental awareness report, created by KNMI and issued to KLM operations, was developed and tested. Given the low frequency of lightning in autumn and winter it is important to make dispatchers and crews aware of the presence of this specific weather scenario. 

Unexpected, sudden lightning strikes in small showers, sometimes triggered by aircraft may lead to unforeseen, additional costs. As induced lightning often occurs at low altitude on approach and descent, the Schiphol Terminal Control Area (TMA) (Figure 4) was chosen as the best area for this test. An awareness report on the phenomena valid for this area only will trigger alertness and may lead to a correct decision for a possible avoidance procedure and so to saving costs.

In the test period (7.1 - 16.4 2008) reports covering all busy hours at the airport were issued daily by the aviation meteorologist in the Central Forecasting Office of KNMI. Figure 5 shows an example of an AIL Awareness Report. By changing the specific colours the meteorologist can highlight specific hours and the spatial frequency of expected showers with a probability of unexpected lightning strikes.

An evaluation of the AIL forecast, using pattern recognition and SAFIR/FLITS data in the Schiphol TMA, gave the following results. Out of 82 days, 48 (58%) days showed a reported wind at 700 hPa in the sector 250-020 degrees. 25 days (of 82) were classified as days with reported lightning, either by pilots or by the SAFIR/FLITS system, 19 of these with a north westerly wind. Obviously not all days with north westerly winds contribute to unstable weather. The combination of criteria on winddirection, low freezing level (<3000ft) and cold upper air (T< -100C at 700 hPa) showed a good result : 68% of the reported lightning strikes were predicted by this method. With a computed Probability of Detection of 86% and a False Alarm Ratio of 24% the value of the report is in par with other aviation meteorology products(Figure 6).

WinterKOUW
Complementary to KOUW (4), an operational probabilistic forecast system of (severe) thunderstorms in the warm season for the purpose of issuing a weather alarm, WinterKOUW was developed for the cold season (5). KOUW is a Model Output Statistics (MOS) system based on a statistical relation between the occurrence of lightning discharges, and a set of potential predictors calculated from the Numerical Weather Prediction models HIRLAM and ECMWF and an ensemble of advected lightning and radar data (6). In KOUW the (conditional) probability of (severe) thunderstorms is calculated for 12 districts shown in Figure 7. As lightning is less frequent in winter, WinterKOUW presents the probability of ≥1 lightning discharge in only 4 districts around Amsterdam Airport (almost in the centre, Figure 7). Out of a large pool of potential predictors such as traditional instability indices, the most useful predictors appeared to be the HIRLAM Boyden index, the ECMWF convective precipitation, two radar advection predictors and the HIRLAM Convective Available Potential Energy (CAPE). 

Figure 7 is an example of a WinterKOUW probability forecast. The forecast probabilities for this example are (much) higher than the climatological probabilities of <=7% (period 2004 – 2007) for the 4 districts. Several lightning reports were received through Air Trafic Control and detected by SAFIR/FLITS. Green area corresponds to 31% and 1 report, yellow 58% , 3 reports, and blue 15% and 9%, no reports. The WinterKOUW forecast system turned out to have a good skill compared to the climatology.

Lightning reports 
From the Fahey study(2) it is known that the number of lightning reports in the USA from aircraft in flight in the cold season exceeds the number of reports in the warm season. For a comparison with the Dutch Airspace, reports of aircraft lightning encounters were obtained from Air Trafic Control for the period May 2005 – April 2008. Knowing that not all encounters are reported, the numbers must be treated as a test at random. In total 33 lightning reports were received. Only 10 were reported in the warm season, of which 8 at unknown altitude, 1 in take off and 1 in descent. For the cold season 23 reports were received, 12 at unknown altitude, 6 in take off and 5 in descent. Comparing these numbers it may be concluded that winter-thunderstorms are en route more difficult to detect or recognize in relation to summer storms.