Central Park Heat Wave Climatology

JUMP TO:

Jump to SectionBackgroundData and MethodsResultsSynoptic CompositesSynoptic ClassificationConclusionsWorks Cited

Jump to Figures 1-29Weather Related FatalitiesHigh Pressure Types to Affect NEUSFrequency of Heat WavesAverage High Temps for Heat WavesBox and Whisker of 3-day Heat Waves for Max TempsBox and Whisker of 4-day Heat Waves for Max TempsBox and Whisker of 5-day Heat Waves for Max TempsAverage Low Temps for Heat WavesBox and Whisker for 3-day Heat Waves for Low TempsBox and Whisker for 4-day Heat Waves for Low TempsBox and Whisker for 5-day Heat Waves for Low TempsDecadal Frequency of Heat WavesNumber of Heat Waves by MonthComposite 850 hPa Temps with Central Park and OKX Labeled3-day Heat Wave 850 hPa Temps4-day Heat Wave 850 hPa Temps5-day Heat Wave 850 hPa TempsMax Hourly TemperaturesDaytime Temperature and Dew Point DifferencesTime Interval That Max Hourly Temps are ReachedObserved Dew Points at Time of Max Hourly TempMax Hourly Heat IndexComposite MSLP for 3-day Heat WavesComposite MSLP for 4-day Heat WavesComposite MSLP for 5-day Heat WavesComposite MSLP Anomaly for 3-day Heat WavesComposite MSLP Anomaly for 4-day Heat WavesComposite MSLP Anomaly for 5-day Heat WavesComposite 500 hPa Geopotential Height for 3-day Heat Waves

Jump to Figures 30-59Composite 500 hPa Geopotential Height for 4-day Heat WavesComposite 500 hPa Geopotential Height for 5-day Heat WavesComposite 500 hPa Geopotential Height Anomaly for 3-day Heat WavesComposite 500 hPa Geopotential Height Anomaly for 4-day Heat WavesComposite 500 hPa Geopotential Height Anomaly for 5-day Heat WavesComposite 850 hPa Temperature for 3-day Heat WavesComposite 850 hPa Temperature for 4-day Heat WavesComposite 850 hPa Temperature for 5-day Heat WavesComposite 850 hPa Temperature Anomaly for 3-day Heat WavesComposite 850 hPa Temperature Anomaly for 4-day Heat WavesComposite 850 hPa Temperature Anomaly for 5-day Heat WavesComposite Wind Speed and Direction for 3-day Heat WavesComposite Wind Speed and Direction for 4-day Heat WavesComposite Wind Speed and Direction for 5-day Heat WavesDistribution of Yarnal TypesComposite MSLP for 3-day Heat Waves for BH typeComposite MSLP for 3-day Heat Waves for PH typeComposite MSLP (CONUS view) for 3-day Heat Waves for PH typeComposite MSLP for 3-day Heat Waves for EH typeComposite MSLP for 3-day Heat Waves for HS typeComposite 500 hPa Geopotential Height for 3-day Heat Waves for BHComposite 500 hPa Geopotential Height for 3-day Heat Waves for PHComposite 500 hPa Geopotential Height for 3-day Heat Waves for EHComposite 500 hPa Geopotential Height for 3-day Heat Waves for HSComposite 850 hPa Temp for 3-day Heat Waves for BHComposite 850 hPa Temp for 3-day Heat Waves for PHComposite 850 hPa Temp for 3-day Heat Waves for EHComposite 850 hPa Temp for 3-day Heat Waves for HSComposite Surface Wind for 3-day Heat Waves for BHComposite Surface Wind for 3-day Heat Waves for PH

Jump to Figures 60-89Composite Surface Wind for 3-day Heat Waves for EHComposite Surface Wind for 3-day Heat Waves for HSBox and Whisker for BH for 3-day Heat WavesBox and Whisker for EH for 3-day Heat WavesBox and Whisker for PH for 3-day Heat WavesBox and Whisker for HS for 3-day Heat WavesBox and Whisker for BH for 4-day Heat WavesBox and Whisker for EH for 4-day Heat WavesBox and Whisker for PH for 4-day Heat WavesBox and Whisker for HS for 4-day Heat WavesBox and Whisker for BH for 5-day Heat WavesBox and Whisker for PH for 5-day Heat WavesBox and Whisker for Daily BH Dew Points for 3-day Heat WavesBox and Whisker for Daily EH Dew Points for 3-day Heat WavesBox and Whisker for Daily PH Dew Points for 3-day Heat WavesBox and Whisker for Daily HS Dew Points for 3-day Heat WavesBox and Whisker for Daily BH Dew Points for 4-day Heat WavesBox and Whisker for Daily EH Dew Points for 4-day Heat WavesBox and Whisker for Daily PH Dew Points for 4-day Heat WavesBox and Whisker for Daily HS Dew Points for 4-day Heat WavesBox and Whisker for Daily BH Dew Points for 5-day Heat WavesBox and Whisker for Daily PH Dew Points for 5-day Heat WavesBox and Whisker for Hourly Temps for BH for 3-day Heat WavesBox and Whisker for Hourly Temps for PH for 3-day Heat WavesBox and Whisker for Hourly Temps for EH for 3-day Heat WavesBox and Whisker for Hourly Temps for HS for 3-day Heat WavesBox and Whisker for Hourly Dew Points for BH for 3-day Heat WavesBox and Whisker for Hourly Dew Points for PH for 3-day Heat WavesBox and Whisker for Hourly Dew Points for EH for 3-day Heat WavesBox and Whisker for Hourly Dew Points for HS for 3-day Heat Waves

Jump to TablesTable1Table2Table3Table4Table5Table6Table7Table8Table9Table10Table11Table12Table13Table14Table15

BACKGROUND

 

Extreme heat is a major weather killer in the United States (Robinson 2001). Data gathered from the National Centers for Environmental Information (NCEI, formerly the National Climatic Data Center) (figure 1) shows that in a 12 year period from 1998 to 2009, heat has caused the most deaths for 5 out of the 12 years. Although this is not more than half of the years, it comes out number one in more years than any other weather related phenomenon. Looking at the 12 year average however clearly shows that heat comes out ahead of all the other phenomena, with an average of 154 deaths, followed by convection with 129. The only time heat is surpassed by a large margin is with Hurricane Katrina in 2005 where over 1000 people were killed. 1999 saw a peak in heat related fatalities, likely due to the extremely hot summer (Pollina et al., 2013).

 

 

Figure 1. Weather related fatalities from 1998-2009.

 

Climate change is a well established scientific fact. The Fifth Assessment Report from the Intergovernmental Panel on Climate Change states that "the science now shows with 95% certainty that human activity is the dominant cause of observed warming since the mid 20th century". Climate models show that there will be anywhere from a 0.5°C to just under 4.5°C increase in global temperatures at the surface by the year 2100 (Cubasch et al. 2013). As temperatures warm in the future, humans will become more susceptible to the effects of extreme heat.

Urban areas are more likely to suffer from extreme heat due to the urban heat island effect, a well known meteorological phenomenon where urban areas are warmer than the surrounding rural areas due to storage of short wave radiation in the concrete, asphalt, and buildings of the urban setting as opposed to the vegetative nature of the rural areas (Solecki et al. 2005, U.S. EPA 2013). Nighttime temperatures show a more drastic difference between urban and rural areas as opposed to daytime temperatures. For example, temperatures on a calm, cloudless night can be up to 20 degrees cooler at Westhampton Beach, NY than in Central Park, NY, which is located 70 miles east of Central Park. Therefore, extreme heat becomes more of a problem for urban areas, especially so given the fact that the global temperature is expected to continue to increase in the future.

The Earth is warming and urban areas will be more at risk for extreme heat, but why even care about heat since we have air conditioning? Air conditioning has become more prevalent in NYC over the decades. In 1970, only about 39% of NYC households had air conditioning. By 1987, that figure increased to 57%. By 2009, close to 88% of all households in NYC has air conditioning (United States Census Bureau). It would seem that with air conditioning becoming more prevalent, we need not worry about heat. That would be true if NYC stood at 100% of all households with air conditioning, but as of 2009, it does not. Some cannot afford it, such as the elderly and those in lower socioeconomic areas. But also, what would happen if a storm Such as Superstorm Sandy occurred in the summer? That could leave potentially millions of people without power and therefore air conditioning, even if 100% of NYC households had air conditioning.

What is a heat wave? There is no standard definition. In 1901, A.T. Burrows defined a heat wave (or as he termed it, hot wave) as 3 or more days in which the maximum temperature in the shade reached or exceeded 90°F. Robinson et al. (2001) notes that there is a lack of rigorous definition, although he does provide a basic definition as "it is an extended period of unusually high atmosphere-related heat stress, which causes temporary modification in lifestyle and which may have adverse health consequences for the affected population." The Glossary of Meteorology from the National Weather Service defines a heat wave as "a period of abnormally and uncomfortably hot and usually humid weather", but does note that it is based on what is normal for a given region of the world. For example, what is warm in Maine is not considered warm in the U.S. Southwest. A locally defined heat wave will be advantageous in many ways. For the public, when a heat wave is forecast, steps can be taken to alleviate heat stress, such as opening cooling centers as is done in NYC when heat advisories or excessive heat warnings are issued. It will also give clarification for the public. In places where there are no locally defined heat waves, the public may be confused as to whether 2 or 3 days constitutes a heat wave. This may not seem like much of a difference, but the prolonged nature of heat waves, which may contain days of excessive heat, can have negative side effects. Epidemiologic analysis conducted by the NYC Department of Health and Mental Hygiene revealed that there is a cumulative impact on consecutive days of extreme heat, with each day adding to the risk of previous days (Metzger 2010). Finally, it would provide a standard definition for local media, especially in a high market area such as NYC, where multiple media outlets could define a heat wave anyway they see fit.

Because extreme heat is a major killer and the effects of heat waves (which need to be locally defined) are cumulative, along with the fact that the global temperature is expected to continue and urban areas are warmer than their surrounding rural areas, a heat wave study of NYC is needed. This heat wave study will answer the following questions:

• What is a heat wave for NYC?
• How many heat waves have affected NYC?
• On average, what are the maximum and minimum temperatures during these heat wave events, and how do they evolve from day to day?
• What is the climatology of heat waves?
  • earliest and latest?
  • distribution of monthly heat waves?
  • decadal distribution?
• How do hourly temperatures during the day evolve during these events?
• How much warmer than the ambient temperature is the heat index?
• What are the synoptic patterns that are most common during these events?

The results from this study could be used by forecasters of the local Weather Forecast Office (WFO) of New York, NY (OKX) to better forecast these events. If forecasters are better able to predict the onset of a heatwave, they will be able to alert the public about an impending heat wave, allowing the public to take the necessary precautions to prepare for the heatwave. It will also allow Emergency Managers (EM) to take precautions to keep the public healthy by opening up cooling centers.

 

DATA AND METHODS

 

This study uses the Central Park observations for its data because out of all 3 stations in NYC (JFK and LaGuardia being the other 2), Central Park has the longest history of data and would therefore better represent the climate of NYC. The daily maximum temperatures throughout the entire data set were examined and any period where the maximum temperature reached or exceeded 90°F for 3 consecutive days or more was determined to be a heatwave. These temperatures were obtained from NCEI and Weather Underground. The period of record for this project is 1876-2011 for a total of 135 years' worth of data. The daily maximums were averaged together for each day of a heat wave to gain an understanding of what types of maximum temperatures we see during these events and how they evolve from a day to day basis is. For example, all the day 1 maximum temperatures were averaged in a 3 day heatwave and compared to day 2 and day 3 to understand the temporal evolution. The same was done for the daily minimum temperatures during these heat waves to gain the same understanding.

3-hourly temperature and dew point was also obtained from NCEI and Weather Underground. Unfortunately, the 3-hourly data set does not span the entire daily maximum data set. 3-hourly data is available from 1965-1982 and from 1996-2011 for a total of 34 year. From there, the heat index (HI) was determined, using the following formula:

HI=c₁+c₂T+c₃R+c₄TR+c₅T^²+c₆R^²+c₇T^²R+c₈TR^²+c₉T^²R^²  (Rothfusz 1990) where:

c₁=-42.379

c₂=2.04901523

c₃=10.14333127

c₄=-0.22475541

c₅=-6.83783

c₆=-5.481717*10^^-2

c₇=1.22874*10^^-3

c₈=8.5282*10^^-4

c₉=1.99*10^^-6

T=temperature

R=Relative Humidity (RH)

The HI was only calculated for ambient temperatures of 80°F or greater. In a similar fashion to the daily maximum, the 3-hourly temperatures, dew points, RHs, and HIs were averaged during the heatwave events to gain an understanding of how these variables evolve throughout the daytime heating.

The National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis (data available from 1948-2017) and the North American Regional Reanalysis (NARR) (data available from 1979-2017) were used to composite the large-scale flow evolution and other variables for the heat wave events. Daily composites were created using the websites https://www.esrl.noaa.gov/psd/data/composites/hour/ and https://www.esrl.noaa.gov/psd/cgi-bin/data/narr/plotday.pl (Kalnay and Coauthors 1996). Identifying these patterns in numerical weather prediction (NWP) models will alert forecasters to the potential for a heatwave.

850 hPa temperatures from OKX soundings (about 60 miles east of NYC) were obtained through the University of Wyoming (data available from 1995-2017). Composites from the aforementioned models show that the 850 hPa temperatures during these events were very similar to the averaged 850 hPa temperatures obtained from the OKX soundings and therefore the 850 hPa temperatures from the soundings were used, since actual observations would likely be more accurate than a model. 850 hPa temperatures are important because they are often used to predict surface temperatures.

Daily surface maps of the heatwave events were downloaded from the Weather Predictions Center and manually inspected to determine what type of surface high pressure is most common during these events. The Yarnal (1993) synoptic classification system was used to classify the high pressure systems. Yarnal identified eight different types of surface pressure patterns over the Northeast United States (NEUS). The “pre-high” synoptic type occurs when surface high pressure builds southeastward over the NEUS (Fig. 2a), which typically occurs after the passage of a cold front. With the surface high centered to the northwest of the NEUS there is usually northwesterly flow. For the “extended high” (Fig. 2b), the center of surface high pressure is directly over the NEUS, with light winds across the region. The “back of high” pattern has the center of the high located just to the east of the East Coast (Fig. 2c), which allows for southwesterly flow over the NEUS region. With “high to the south” (Fig. 2d), there is also typically a corresponding area of lower pressure to the north, with surface westerly flow between the two pressure centers. On the other hand, a high to the north and a corresponding low to the south allows for an easterly flow across the NEUS (Fig. 2e).

 

Figure 2a-e. High pressure types to affect the NEUS.

 

 

RESULTS

 

There were 263 separate heatwave events to affect Central Park between 1876 and 2011 (Fig. 3). This works out to be just under 2 per year (135 years*2=270). There were 134 3-day heatwave events, which is just about 1 per year. 4-day heatwaves occur a little under half as much as 3-day events, with a total of 63 events, or about once every other year. 5-day heat waves occur just over half as much as 4-day and a quarter as much as 3-day events, with a total of 36 events. That works out to be once every 3.75 years. 6-day heat waves and higher are not statistically significant and occur more randomly, so results from here on out will largely reflect 3-, 4-, and 5-day heat waves. The earliest a heatwave has occurred is April 16-18 2002 (a close 2nd was April 17-19 1976). The latest a heatwave has occurred is September 21-23 1914. The highest order heatwave to ever occur in Central Park is 12, from August 24 to September 4 1953.

 

 

Figure 3. Frequency of heat waves to affect Central Park from 1876-2011.

 

This climatology is based on maximum temperatures. Therefore, the maximum temperatures for each day of each heat wave type (3-day, 4-day, 5-day, etc.) was taken and averaged. The results are shown in figure 4. 3-day heat waves start out with high temperatures in the lower 90s (92.3ºF to be exact), then warm to the lower to middle 90s (93.5ºF), then cool slightly (93ºF). The 4 and 5-day heat waves follow a similar pattern of warming and then cooling. 4-day heat waves warm through day 3 and then cool on the last day of a heat wave. 4-day heat waves start off in the lower 90s (92.6ºF) then gradually warm to the middle 90s by day 3 (94.9ºF), then cool to the lower 90s by day 4 (93.1ºF). 5-day heat waves start off in the lower 90s (91.9ºF), then warm into the middle to upper 90s (96ºF), then cool to the lower to middle 90s by day 5 (93.8ºF). High temperatures for 4-day heat waves are consistently warmer than 3-day heat waves (comparing days 1, 2, and 3). Except for day 1, 5-day heat waves are warmer than 3-day heat waves (comparing days 1, 2, and 3) and 4-day heat waves (comparing days 1, 2, 3, and 4). This graph has great forecasting value. After looking at this plot, a forecaster may want to forecast slightly higher temperatures on day 2 of a 3-day heat wave compared to day 1, since, on average day 2 is warmer than day 1. If a 3-day heat wave is extended to become a 4-day heat wave early in the forecast period, temperatures may need to be increased since on average, 4-day heat waves are warmer than 3-day heat waves.

 

 

Figure 4. Average high temperatures for each day of 3-day, 4-day, and 5-day heat waves.

 

Table 1 shows the statistics for maximum (high) temperatures of each day a 3-day heat wave. The minimum high temperatures (which are set to 90ºF, since this is the definition of a heat wave), 25th percentile (meaning that 75% of all observations are above this value), median (the midpoint of all the observations), 75th percentile (meaning that 25% of all observations are above this value), maximum high temperatures, and standard deviation were all calculated for all the high temperatures of each day. Box and whisker diagrams were also created (figure 5) to help visualize the range of values. In figure 5, the lower end whisker is the minimum high. The lower portion of the red box represents the 25th percentile. The median is the small dashed line inside the red box. The top of the red box represents the 75th percentile. Finally, the top whisker represents the highest observation recorded. All these variables follow the trend of the averages, that being a slight increase in the temperatures between days 1 and 2, then a decrease from days 2 to 3.

 

 

Table 1. Minimum, 25th percentile, median, 75th percentile, maximum, and standard deviation for high temperatures of each day of a 3-day heat wave.

Day	Minimum High	25th Percentile	Median	75th Percentile	Maximum High	Standard Deviation
1	90	91	92	93	101	2.4
2	90	92	93	95	103	2.7
3	90	91	92	94.75	102	2.6

 

 

Figure 5. Box and whisker diagrams for days 1-3 of 3-day heat waves for maximum temperatures. The bottom whisker is the lowest maximum temperature recorded on each day (all 3 are 90ºF since this is the lowest possible temperature given our definition of a heatwave). The bottom portion of the red box is the 25th percentile, the small dashed line inside the red box is the median, the top portion of the red box is the 75th percentile, and the top whisker is the highest maximum temperature recorded.

 

 

Table 2 shows the statistics for high temperatures of each day a 4-day heat wave and figure 6 is the respective box and whisker diagrams. 

 

Table 2. Same as in Table 1 except for 4-day heat waves.

Day	Minimum High	25th Percentile	Median	75th Percentile	Maximum High	Standard Deviation
1	90	91	92	94	101	2.4
2	90	92	93	96	104	3.1
3	90	93	95	97	103	3.1
4	90	91	93	95	100	2.5

 

 

 

Figure 6. Same as figure 5, except for 4-day heat waves.

 

Table 3 shows the statistics for high temperatures of each day a 5-day heat wave and figure 7 is the respective box and whisker diagrams. 

Table 3. Same as in Table 1, except for 5-day heat waves.

Day	Minimum High	25th Percentile	Median	75th Percentile	Maximum Percentile	Standard Deviation
1	90	90	91	93	101	2.3
2	90	92	94	96	103	3.0
3	90	93	96	98.25	106	3.8
4	90	93	94.5	97.25	102	3.3
5	90	92	94	95.25	101	3.0

 

 

 

Figure 7. Same as Figure 5 except for 5-day heat waves

 

Average low temperatures (fig. 8) follow a similar pattern with a steady increase in temperatures through each type of heat wave, and then cooling toward the end. 3-day heat waves start out with lows in the lower 70s on average (70.8ºF) for day 1, then warm into the middle 70s (74.2ºF) for day 2, then cool slightly on the 3rd day (73.9ºF). Average low temperatures on 4-day heat waves start out in the lower 70s as well. In fact, low temperatures on the first day of a 4-day heat wave are exactly the same as the first day of a 3-day heat wave (70.8ºF). Temperatures steadily warm on days 2 and 3 (74.5ºF and 75.3ºF respectively), and then cool on day 4 (73.8ºF). 5-day heat waves start out almost 2ºF cooler than 3 and 4-day heat waves with lows in the upper 60s on average (68.7ºF on day 1), then warm into the middle 70s from days 2 through 4 (73.4ºF, 75.1ºF, and 75.1ºF respectively), with cooler temperatures on day 5 (73.4ºF). Low temperatures on 5-day heat waves are not only cooler for the 1st day, but 2nd and 3rd days as compared to 3 and 4-day heat waves. It is not until day 4 that 5-day low temperatures become warmer than 4-day heat waves.

 

 

Figure 8.  Same as in figure 4, but for low temperatures.

 

The 3-day heat wave low temperature statistics are shown in Table 3 and box and whisker diagrams are shown in figure 9. Since the criteria for a heat wave is based on high temperatures, there is no lower limit for low temperatures. 

Table 4. Same as in Table 1, except for low temperatures.

Day	Minimum Low	25th Percentile	Median	75th Percentile	Maximum Low	Standard Deviation
1	58	68	71	74	80	4.1
2	66	72	74	77	82	3.4
3	62	72	74	76	84	3.5

 

 

Figure 9. Same as in Figure 5, except for low temperatures. Note: there is no set minimum temperature for lows, since heatwaves in this study are based on high temperatures. 

Table 5. Same as in Table 4, except for 4-day heatwaves.

Day	Minimum High	25th Percentile	Median	75th Percentile	Maximum High	Standard Deviation
1	61	68	71	73	79	3.7
2	68	73	74	76	84	3.1
3	68	73	75	77	83	3.4
4	60	72	74	77	84	4.3

 

 

 

Figure 10. Same as in figure 9, except for 4-day heatwaves.

Table 6. Same as in Table 4, except for 5-day heatwaves.

Day	Minimum High	25th Percentile	Median	75th Percentile	Maximum Percentile	Standard Deviation
1	60	67	68.5	71.25	77	3.8
2	67	70	74	76	80	3.4
3	69	72	75.5	77	82	3.1
4	65	74	75.5	77	79	2.7
5	63	71	73	77	80	4.3

 

 

Figure 11. Same as in figure 9, except for 5-day heatwaves.

The next plot (Fig. 12) shows the distribution of each heatwave event through the decades, beginning in the 1880s (although data began in 1876, it was not a complete decade, and was not included in this plot). One of the key things to note with this plot is that prior to the 1930s, rarely do we see anything higher than a 5-day heatwave event. There were two 6-day heatwaves in the 1890s and one 7-day heatwave in the 1900s, but other than that, 3-, 4-, and 5-day heatwaves dominated. From the 1940s and onward, higher order heatwaves become more prevalent. From the 1880s to the 1920s, there is a peak in 3-day heatwaves in the 1890s (occurring 9 times that decade). From the 1930s to the 1990s, 3-day heat waves occur at least as much as the peak of 9 in the 1880-1920 period with a peak in 3-day heat waves occurring in the 1930s with 15 events. 3-day heat waves drop below this peak in the 2000s, with a total of only 7, but there were more higher order events (two 6-day and one each of 8 and 9-day events). Along with an increase in higher order events from the 1930s-2000s is an increase in 4-day heatwaves (outside of the peak of 9 in the 1940s). This plot may reflect the changes occurring with climate change. The increase in 3-day heatwaves beginning in the 1930s and the prevalance of higher order heat waves after the 1940s at least shows more 90+ days and longer duration events.

 

 

 

Figure 12. Number of heatwave events per decade for 3-day through 12-day heatwave events. 

 

On the first and second day of a 3-day heat wave, NYC lies within an area of zero MSLP anomaly. On the first day, a large area of +1 to +2 mb positive anomaly exists over the Mid-Atlantic and Southeast US (fig. 26). The northern extent is just to the south of Long Island and stretches southward to northern Florida, while the western extent is just east of the Mississippi River in Western Kentucky and Tennessee and the eastern extent is hundreds of miles off the east coast in the Western Atlantic Ocean. There is a positive anomaly maximum (+2 to +3 mb above normal) over northern South Carolina, much of North Carolina, and stretching northeastward just off the Virginia coast. There is a large area of negative anomaly over the Mid-West, Plains and eastern Rockies states, with a maximum negative anomaly (-5 to -4 mb below normal) over northeast Colorado, northwest Kansas, central Nebraska and extreme central South Dakota. On day 2, the positive anomaly area sinks south and becomes less intense (closer to neutral) as the negative anomaly moves east, closer to NYC and also weakens. By day 3, the eastern and southern edges of the negative anomaly is over NYC (-1 to -2 mb), while the positive anomaly is mainly limited to just off the South Carolina coast.

 

 

 

 

 

 

 

Figure 31 shows the 500 mb heights for 5-day heat waves. There is a 500 mb trough off the Northeast U.S. coast, a ridge over the Western Great Lakes area, and a trough over the West Coast of the U.S. A broad area of high pressure exists over the Gulf states, northern portions of the Gulf of Mexico and extending east into the Western Central Atlantic. NYC lies on the eastern edge of the ridge, lying between the 5850 m and 5825 m geopotential heights.

On the second day, the trough off the Northeast U.S. moves east and amplifies somewhat, while the ridge over the Western Great Lakes moves into the Central Great Lakes Area with no appreciable change in amplitude. The trough over the West Coast remains nearly stationary and deamplifies. The broad area of high pressure over the Southeast U.S. becomes closed off over the Southern U.S. On the third day, the trough off the Northeast coast tracks slightly east with no appreciable change in amplitude, the ridge weakens, while the trough over the West Coast retrogrades to the west. On the fourth day, the trough off the Northeast U.S. remains nearly stationary while weakening, the ridge over the Great Lakes continues to weaken, now becoming zonal flow over that region, while the trough over the West Coast continues to remain stationary. Further weakening (deamplification) of the trough over the Northeast Coast is seen on the last day, while zonal flow is evident over the Great Lakes and Northeast U.S., and the West Coast trough moves slightly closer back to the West Coast. The closed high over the South retrogrades to the west, situated over Texas, the U.S. Southwest, and northern Mexico.

 

 

Synoptic composites for each Yarnal type were created for various elements for 3, 4, and 5-day heat waves. Since the results were similar for each type of heat wave within a synoptic type, only 3-day heat wave results will be shown, with differences noted.

 

In figure 45, the composite for the 3-day BH type is shown. There is a 1022-1024 mb high pressure centered just west of the Azores which stretches west into the East Coast of United States. The 1016 mb line is closest to NYC. Within the periphery of this high, there is a trough, whose axis extends from southern New England into the mid-Atlantic. The major difference with the other heat waves is that 4-day heat waves have a slightly stronger high (1024-1026 mb) and 5-day heat waves have a weaker trough.

 

 

Figure 45. Composite MSLP for 3-day heat waves for BH type.

The PH synoptic type is shown in fig. 46. It shows a 1022 mb to 1024 mb high pressure over the Azores which stretches west into the northeast Caribbean. There is a sharp surface trough over the NEUS, it is much more evident as compared to the BH type. This could be a composite of the passage of a cold front, which typically precedes the PH type. A closer view of this synoptic type in figure 47 shows that this trough is just off the Northeast Coast, while another closed high (1017 to 1018 mb) is centered over southwest Virginia, western Kentucky, and northeast Tennessee.

 

 

Figure 46. Same as fig. 45, but for PH.

 

 

 

Figure 47. Same as fig. 46, but for CONUS view.

 

The EH MSLP is shown below in fig. 48, which shows a 1023 mb to 1025 mb high centered to the southwest or west-southwest of the Azores. A surface trough is evident over the NEUS. However, this is not nearly as sharply defined as the PH type. Similar to PH, there is another closed high over western Virginia and western North Carolina (1018.5 mb to 1020 mb) .

 

 

Figure 48. Same as fig. 45, but for EH.

 

HS MSLP is shown below (fig. 49). It shows a 1022 mb to 1024 mb high centered just to the south-southwest of the Azores. As with the other synoptic highs, a trough is over the NEUS. Although there is no closed high south of the NEUS (which may be due to averaging out the data), there is a low pressure situated over Newfoundland in Canada which is typical of the HS type. This low is between 1004 mb and 1006 mb in strength.

 

 

Figure 49. Same as fig. 45, but for HS.

For the BH 500 mb heights (fig. 50), there is a 5900 m closed high to the south of the area. This closed high is off the North Carolina/South Carolina coasts for 3-day heat waves. It is broader for 4- and 5-day heat waves, extending westward into the Southeast U.S. for 4-day heat waves, and into all of the Southern U.S. for 5-day heat waves. There is weak ridging in the vicinity of the NEUS. On 3-day heat waves, it is over Eastern New York State. This ridge is progressively farther east for 4- and then 5-day heat waves, where it is just offshore. The 5875 m line is closest to NYC for all 3 heat wave types. This is the highest height to affect NEUS out of the main synoptic types to affect the region during heat waves.

 

 

 

 

Figure 50. Composite 500 mb geopotential height for 3-day heat waves for BH type.

For 500 mb PH heights (fig. 51), there is a large 5900 m closed high over the southern Plains states and into northern Mexico. Over the NEUS, a weak trough is just off the NEUS coast. The 5850 mb line is closest to NYC.

 

 

Figure 51. Same as fig. 50, but for PH.

 

The EH 500 mb heights are shown in fig. 52. This shows the NEUS in between an exiting trough well off the NEUS coast and entering ridge over the eastern Great Lakes region. A 5900 m closed high is centered over western Virginia/North Carolina and eastern Kentucky/Tennessee. The 5875 m height is closest to NYC

 

Figure 52. Same as fig. 50, but for EH.

The HS 500 mb height is shown below (fig. 53). This shows a zonal flow over the NEUS, with a semi-closed high over the Southeast US. The 5825 m height is closest to NYC. HS shows the lowest heights to affect NYC.

 

 

Figure 53. Same as fig. 50, but for HS.

Composite 850 mb temperatures for BH, PH, EH, and HS are shown in fig. 54 - 57 respectively. For BH, 850 mb temperature for NYC is 17°C to 18°C. For PH, it is 16°C to 17°C. For EH, 850 mb temperature is 17°C to 18°C, and for HS it is 16.5°C-17.5°C.

 

 

 

Figure 54. Composite 850 mb temperature for 3-day heat waves for BH type.

 

 

 

Figure 55. Same as fig. 54 but for PH.

 

 

 

 

Figure 56. Same as fig. 54, but for EH. 

 

 

 

 

Figure 57. Same as fig. 54, but for HS.

The surface winds composite are shown in figures 58- 61 (BH, PH, EH, and HS respectively). The closest vector to NYC is over west central New Jersey. BH shows a composite southwest wind of 4.0 to 5.0 m/s (8.9 to 11.2 mph) over NYC, with the highest winds just east of Long Island at 6 to 6.5 m/s (13.4 to 14.5 mph), while PH shows that NYC is in between an area of west wind, with an average wind speed of 3.5 m/s to 4.0 m/s (7.8 to 8.9 mph). For EH, a composite southwest flow occurs with an average wind speed of 3.0 m/s to 3.5 m/s (6.7 to 7.8mph). Finally, HS shows a southwest wind at an average speed of 4.5 to 5.5 m/s (10.0 to 12.3 mph), with the highest winds similar to BH just east of Long Island at 7.0 to 7.5 m/s (15.6 to 16.8 mph). Of all 4 synoptic types, HS and BH show the highest wind speeds, while EH and PH show the lowest. It stands to reason that EH would show the lowest, since high pressure should be centered over the region with this synoptic type. With a lack of pressure gradient, winds will be lower.

 

 

 

Figure 58. Composite surface wind speed (shaded) and direction (arrows) for 3-day heat waves for BH type.

 

 

 

 

Figure 59. Same as fig. 58, but for PH. 

 

 

 

 

Figure 60. Same as fig. 58, but for EH.

 

 

 

 

Figure 61. Same as fig. 58, but for HS.

Separate statistics for high temperatures, low temperatures and dew points for the synoptic classes were computed in the form of box and whiskers diagrams. These diagrams show statistics for the Yarnal synoptic types for each heat wave.

For 3-day heat waves, tables 7 and 8 show the statistics for the main Yarnal types that cause 3-day heat waves for highs and lows respectively, while fig. 62-65 show box and whisker diagrams for each synoptic type. Notice that the minimum high temperatures of 90ºF are all the same. This is because a heat wave is defined by this minimum temperature. The EH type shows the greatest high temperature of all synoptic types that cause a 3-day heat wave at 103ºF, with PH the lowest at 99ºF. EH also has the highest variability with a standard deviation of 2.9, while the lowest occurs with PH at 2.4. BH has a standard deviation of 2.6 and HS has a standard deviation of 2.7. The median is 92ºF for all types except HS, where it is 93ºF. The 25^th percentile is 91ºF for all types, again, except for HS, where it is 92ºF. The 75^th percentile is 94ºF for BH and PH; it is 95ºF for EH and HS.

 

For low temperatures, the maximum lows are either 82ºF for BH and PH or 84ºF for EH and HS. The lowest minimum low temperature occurs with PH at 58ºF and the highest with BH at 62ºF. EH has a minimum low of 60ºF, while it is 61ºF for HS. EH has the lowest median temperature of 72ºF, while it is 73ºF for PH, 74ºF for BH and 75ºF for HS. The 25^th percentiles start at a low of 70ºF with EH, BH and PH are 71ºF and HS is 72ºF. The 75^th percentile are 74ºF, 75ºF, 76ºF, 77ºF for EH, PH, BH, and HS respectively. Finally, standard deviations range from a low (least variability) in EH at 3.7, to a high (most variability) in PH at 4.8. HS and BH have similar standard deviations at 4.0 and 4.1 respectively.

 

Table 7. Maximum, minimum, median, 25th percentile, 75ht percentile, and standard deviation for BH, EH, PH, and HS, for 3-day heat waves.

Yarnal type	Maximum high temp	Minimum high temp	Median	25th percentile	75th percentile	Standard deviation
BH	102	90	92	91	94	2.6
EH	103	90	92	91	95	2.9
PH	99	90	92	91	94	2.4
HS	102	90	93	92	95	2.7

 

Table 8. Same as table 7, but for low temperatures.

Yarnal type	Maximum low temp	Minimum low temp	Median	25th percentile	75th percentile	Standard deviation
BH	82	62	74	71	76	4.1
EH	84	60	72	70	74	3.7
PH	82	58	73	71	75	4.8
HS	84	61	75	72	77	4.0

 

 

 

Figure 62. Box and whisker diagrams for high and low temperatures for 3-day heat waves for BH type.

 

 

 

Figure 63. Same as fig. 62, but for EH.

 

 

 

Figure 64. Same as fig. 62, but for PH.

 

 

Figure 65. Same as fig. 62, but for HS.

 

For 4-day heat waves, the greatest maximum temperature occurs with BH at 104ºF (table 9), the lowest maximum high temperature is with EH at 97ºF. HS has a maximum high temperature of 99ºF, while PH has a maximum of 100ºF. Median temperatures are the same for BH, EH, and PH at 93ºF, while it is slightly higher for HS at 94ºF. The 25^th percentile is 90ºF, 91ºF, 92ºF, and 93ºF for EH, BH, PH, and HS respectively. The 75^th percentile for EH is 94ºF, while it is 95ºF for both PH and HS, and 96ºF for BH. The standard deviation is 2.3 for EH and HS, 2.6 for PH and it is the highest for BH at 3.2.

 

For low temperatures for 4-day heat waves, the greatest maximum low temperature is from BH with 84ºF (table 10), while the lowest maximum low temperature is 81ºF from HS. The greatest minimum low temperature occurs with EH at 68ºF, while lowest minimum low temperature occurs with BH at 61ºF. The highest median occurs with HS at 76ºF, then it is 74ºF, 73ºF, 72, with BH, PH, and EH respectively. The 25^th percentile is 71ºF for EH, PH, and HS, while it is slightly higher for BH at 72ºF. HS has the highest 75^th percentile at 77ºF, then it is 76ºF for BH and 75ºF for both EH and PH. EH has the lowest standard deviation with 3.0, then PH with 3.6. HS and BH have similar standard deviations of 3.9 and 4.0 respectively.

Figures 66-69 show the box and whisker diagrams for BH, EH, PH, and HS respectively.

Table 9. Same as table 7, but for 4-day heat waves.

Yarnal type	Maximum high temp	Minimum high temp	Median	25th percentile	75th percentile	Standard deviation
BH	104	90	93	91	96	3.2
EH	97	90	93	90	94	2.3
PH	100	90	93	92	95	2.6
HS	99	90	94	93	95	2.3

 

 

Table 10. Same as table 8, but for 4-day heat waves. 

Yarnal type	Maximum low temp	Minimum low temp	Median	25th percentile	75th percentile	Standard deviation
BH	84	61	74	72	76	4.0
EH	78	68	72	71	75	3.0
PH	82	63	73	71	75	3.6
HS	81	66	76	71	77	3.9

 

 

 

Figure 66. Same as fig. 62, but for 4-day heat waves.

 

 

 

Figure 67. Same as fig. 63, but for 4-day heat waves.

 

 

Figure 68. Same as fig. 65, but for 4-day heat waves. 

 

 

Figure 69. Same as fig. 66, but for 4-day heat waves. 

For 5 day heat wave, EH and HS will be missing, since there was not enough data to produce meaningful results from these synoptic types. The greatest maximum high temperature occurs with PH at 103ºF (table 11), while it is 101ºF for BH. The median for both BH and PH are 94ºF. The 25^th percentiles are at 91ºF and 93ºF for PH and BH respectively. The 75^th percentile is 96ºF and 96.5ºF for PH and BH respectively. Finally, the standard deviation is 2.9 and 3.2 for BH and PH respectively.

 

The greatest maximum low temperature occurs with both BH and PH at 80ºF (table 12). PH as the lowest minimum temperature at 60ºF, while it is 66ºF for BH. The median for both PH and BH is 74ºF. The 25^th percentile is 68.5ºF and 71ºF for PH and BH respectively. The 75^th percentile is 76ºF and 77ºF for PH and BH respectively. The standard deviation is 3.5 and 4.8 for BH and PH respectively.

Figures 70-71 show the box and whisker diagrams for BH and PH respectively.

Table 11. Same as table 7, but for 5-day heat waves. There was not enough data to produce meaningful results for EH and HS.   

Yarnal type	Maximum high temp	Minimum high temp	Median	25th percentile	75th percentile	Standard deviation
BH	101	90	94	93	96.5	2.9
EH
PH	103	90	94	91	96	3.2
HS

 

Table 12. Same as table 8, but for 5-day heat waves. There was not enough data to produce meaningful results for EH and HS.

Yarnal type	Maximum low temp	Minimum low temp	Median	25th percentile	75th percentile	Standard deviation
BH	80	66	74	71	77	3.5
EH
PH	80	60	74	68.5	76	4.8
HS

 

 

 

 

Figure 70. Same as fig. 62, but for 5-day heat waves.

 

 

Figure 71. Same as fig. 64, but for 5-day heat waves.

Table 13 shows the statistics for daily 3-day Yarnal dew points, while figs. 73-74 show the box and whisker diagrams. BH recorded the highest dew point at 77ºF (tied with HS). While it also recorded the lowest dew point at 48ºF, it recorded the highest median, 25th percentile, and 75th percentile at 68ºF, 64ºF, and 71ºF respectively. The fact that for a majority of the statistics regarding dew points, it came out as the highest makes sense. BH is typically a Bermuda high type that introduces a south or southwesterly flow in the region. This means a more humid air mass, and thus typically higher dew points than the other Yarnal types. A possible explanation for BH having the lowest dew point is that it may have been recorded during a transitioning event. Typically, the NEUS will experience a transition of Yarnal types: first PH with high pressure to the west of the area, then EH with high pressure over the area, and finally BH with high pressure offshore, and then back to PH again after the passage of a cold front, repeating the cycle (Pollina et al). The minimum dew point may have been recorded during a transition period from EH to BH, or from BH to PH after the passage of cold front, which would allow high dew points to lower as a drier air mass filters into the region.

 

PH is typically a drier air mass, since it introduces a northwest flow (off shore). It would therefore be expected to show lower dew points during heat wave events. While the maximum and minimum are not the lowest among the Yarnal types (74ºF and 51ºF respectively), the median, 25th percentile, and 75th percentile are (63ºF, 58ºF, and 67ºF respectively). The median, 25th percentile and 75th percentile are a more consistent way to show that PH is a drier air mass, since they take into account all values, not just one (such as the case with the maximum and minimum). 

BH and PH show the highest variability with standard deviations of 5.6. This could hint at the fact that the transitioning between synoptic types causes a high amount of variability for these types. EH has the lowest variability among the synoptic types at 4.6.

Figs. 72-75 show box and whisker diagrams for BH, EH, PH, and HS respectively.

Table 13. Maximum, minimum, median, 25th percentile, 75th percentile, and standard deviation of daily dew points for each BH, EH, PH, and HS for 3-day heat waves.

Yarnal type	Maximum dew point	Minimum dew point	Median	25th percentile	75th percentile	Standard deviation
BH	77	48	68	64	71	5.6
EH	73	53	65	62	69	4.6
PH	74	51	63	58	67	5.6
HS	77	55	64	62	68.5	5.3

 

 

 

 

 

 

 

Back to Top

 

 

Figure 72. Box and whisker diagram for daily BH dew points.

 

 

 

Figure 73. Same as fig. 72, but for EH.

 

 

Figure74. Same as fig. 72, but for PH.

 

 

Figure 75. Same as fig. 72, but for HS.

 

With 4-day daily synoptic type dew points in table 14, BH shows the highest maximum dew point out of all the synoptic types at 77ºF. However, like 3-day heat waves, it comes in at the lowest minimum dew point out of all synoptic types at 48ºF. Again, this may be due to transitioning synoptic types. Median and 25th percentile are highest out of all the synoptic type (69ºF and 67ºF respectively), however, it is only the 2nd highest at 75th percentile (72ºF), coming in behind HS. Although winds are typically westerly with an HS type, a slight turn to a more southwesterly flow would introduce a more humid air mass than would otherwise be seen in an HS type, which also can be seen in a transitory state from an HS to a BH.

 

PH for shows some evidence of being a drier air mass in 4-day heat waves. It shows the lowest dew points for the median, 25th percentile, and 75th percentile (62.5ºF, 60.5ºF, and 70ºF respectively) out of all the synoptic types. EH shows the lowest maximum dew point at 72ºF and as previously mentioned, BH shows the lowest minimum dew point. Again, these may occur during transitory states. 

EH once again shows the lowest variability in 4-day heat waves, with a standard deviation of 4.6, while PH again shows the highest at 5.2. 

 

Table 14. Same as in table 13, but for 4-day heat waves.

Yarnal type	Maximum dew point	Minimum dew point	Median	25th percentile	75th percentile	Standard deviation
BH	77	48	69	67	72	4.9
EH	72	58	65.5	60.5	70	4.6
PH	73	51	62.5	58	65	5.2
HS	75	61	67	62	73	5.1

 

 

 

 

Figure 76. Same as fig. 72, but for 4-day heat waves.

 

 

 

Figure 77. Same as in fig. 73, but for 4-day heat waves.

 

 

 

Figure 78. Same as in fig. 74, but for 4-day heat waves.

 

 

 

Figure 79. Same as in fig. 75, but for 4-day heat waves. 

As with temperatures, EH and HS will be missing for 5-day daily synoptic dew points, since there was not enough data to produce meaningful results from these synoptic types. BH comes in with the higher dew points for each category, while PH therefore come in at the lowest. 

 

Table 15. Same as in table 13, but for 5-day heat waves.

Yarnal type	Maximum dew point	Minimum dew point	Median	25th percentile	75th percentile	Standard deviation
BH	76	57	68	66	70	3.7
EH
PH	75	51	60	57	65	5.5
HS

 

 

 

 

Figure 80. Same as in fig. 72, but for 5-day heat waves.

 

 

 

Figure 81. Same as in fig. 74, but for 5-day heat waves.

 

Box and whisker diagrams were also created for hourly temperatures for each Yarnal synoptic type for 3-, 4-, and 5-day heat waves. 3-day heat waves had the biggest sample size, so they are shown here. 4- and 5-day heat waves did not have enough data to create meaningful plots.

 

For the BH type (fig. 82), the median temperature increases from 78ºF at 12Z to 86ºF at 15Z to 91ºF at 18Z and then decreases down to 90ºF by 21Z.

12Z temperatures had a range from a low of 65ºF to a high of 88ºF. The 25th percentile was at 75ºF, while the 75th percentile was at 80ºF. The standard deviation was 4.23.

15Z temperatures had a range from a low of 76ºF to a high of 97ºF. The 25th percentile was 83ºF, while the 75th percentile was 89ºF. The standard deviation was 3.86.

18Z temperatures had a range from a low of 85ºF to a high of 103ºF. The 25th percentile was 89ºF, while the 75th percentile was 94ºF. The standard deviation was 3.47.

Finally, 21Z temperatures had a range from a low of 76ºF to a high of 101ºF. The 25th percentile was 88ºF, while the 75th percentile was 93ºF. The standard deviation was 4.68.

 

 

Figure 82. Box and whisker diagrams for hourly temperatures for BH synoptic type for 3-day heat waves.

 

For the PH type (fig. 83), the median temperature increases from 76ºF at 12Z to 84 at 15Z to 90ºF by 18Z and then continues to increase to 91ºF by 21Z.

12Z temperatures had a range from a low of 69ºF to a high of 84ºF. The 25th percentile was at 73ºF, while the 75th percentile was at 78ºF. The standard deviation was 3.73.

15Z temperatures had a range from a low of 79ºF to a high of 92ºF. The 25th percentile was 82ºF, while the 75th percentile was 87ºF. The standard deviation was 3.16.

18Z temperatures had a range from a low of 85ºF to a high of 96ºF The 25th percentile was 89.25ºF, while the 75th percentile was 92ºF. The standard deviation was 2.64.

Finally, 21Z temperatures had a range from a low of 83ºF to a high of 96ºF. The 25th percentile was 90ºF, while the 75th percentile was 92ºF. The standard deviation was 2.84.

 

 

Figure 83. Same as in fig. 82, but for PH type. 

 

For the EH type (fig. 84), the median temperature increases from 76ºF at 12Z to 85 at 15Z to 91ºF by 18Z and then decreases to 90ºF by 21Z.

12Z temperatures had a range from a low of 68ºF to a high of 90. The 25th percentile was at 74ºF, while the 75th percentile was at 80ºF. The standard deviation was 4.50.

15Z temperatures had a range from a low of 72ºF to a high of 95. The 25th percentile was 83.2ºF5, while the 75th percentile was 88.75ºF. The standard deviation was 4.38.

18Z temperatures had a range from a low of 82ºF to a high of 100. The 25th percentile was 89ºF, while the 75th percentile was 93.25ºF. The standard deviation was 3.73.

Finally, 21Z temperatures had a range from a low of 83ºF to a high of 102. The 25th percentile was 88ºF, while the 75th percentile was 92ºF. The standard deviation was 4.40.

 

 

Figure 84. Same as in fig. 82, but for EH type.

 

For the HS type (fig. 85), the median temperature increases from 77ºF at 12Z to 84.5 at 15Z to 90.5ºF by 18Z and then decreases to 90ºF by 21Z.

12Z temperatures had a range from a low of 66ºF to a high of 81ºF. The 25th percentile was at 72ºF, while the 75th percentile was at 78ºF. The standard deviation was 4.62.

15Z temperatures had a range from a low of 75ºF to a high of 89ºF. The 25th percentile was 82.75ºF, while the 75th percentile was 86.25ºF. The standard deviation was 3.75.

18Z temperatures had a range from a low of 86ºF to a high of 95ºF. The 25th percentile was 89.75ºF, while the 75th percentile was 92.25ºF. The standard deviation was 2.37.

Finally, 21Z temperatures had a range from a low of 83ºF to a high of 94ºF. The 25th percentile was 85.75ºF, while the 75th percentile was 93.25ºF The standard deviation was 4.08.

 

 

Figure 85. Same as in fig. 82, but for HS type.

 

Box and whisker diagrams were also created for hourly dew points for each Yarnal synoptic type for 3-day heat waves. As with temperatures, since they were the most common type, 3-day heat waves were the only heat waves that hourly box and whisker diagrams were created. 4- and 5-day heat waves did not have enough data to create meaningful plots.

For the BH type (fig. 86), the median dew point is 68ºF at 12Z, 15Z and 18Z,  then decreases to 67ºF by 21Z.

12Z dew points had a range from a low of 49ºF to a high of 75ºF. The 25th percentile was at 64ºF, while the 75th percentile was at 70ºF. The standard deviation was 5.39.

15Z dew points had a range from a low of 52ºF to a high of 76ºF. The 25th percentile was 65ºF, while the 75th percentile was 77ºF. The standard deviation was 5.67.

18Z dew points had a range from a low of 52ºF to a high of 76ºF. The 25th percentile was 62ºF, while the 75th percentile was 70ºF. The standard deviation was 5.96.

Finally, 21Z dew points had a range from a low of 48ºF to a high of 76ºF. The 25th percentile was 61ºF, while the 75th percentile was 70ºF. The standard deviation was 5.78.

 

 

Figure 86. Box and whisker diagrams for hourly dew points for BH type for 3-day heat waves.

 

For the PH type (fig. 87), the median dew point increases from 62ºF at 12Z to 62.5ºF at 15Z decreases back to 62ºF by 18Z and then continues to decrease to 61ºF by 21Z.

12Z dew points had a range from a low of 55ºF to a high of 73ºF. The 25th percentile was at 58ºF, while the 75th percentile was at 64.75ºF. The standard deviation was 4.99.

15Z dew points had a range from a low of 55ºF to a high of 74ºF. The 25th percentile was 58.75ºF, while the 75th percentile was 65ºF. The standard deviation was 5.01.

18Z dew points had a range from a low of 52ºF to a high of 74ºF The 25th percentile was 58ºF, while the 75th percentile was 65ºF. The standard deviation was 5.58. 

Finally, 21Z dew points had a range from a low of 51ºF to a high of 73ºF. The 25th percentile was 57ºF, while the 75th percentile was 65.5ºF. The standard deviation was 6.00. 

 

 

Figure 87. Same as in fig. 86, but for PH type. 

 

For the EH type (fig. 88), the median dew point decreases from 66ºF at 12Z to 65.5ºF at 15Z to 65ºF by 18Z and then increases to 67ºF by 21Z.

12Z dew points had a range from a low of 55ºF to a high of 70ºF. The 25th percentile was at 62.25ºF, while the 75th percentile was at 68ºF. The standard deviation was 4.19.

15Z dew points had a range from a low of 54ºF to a high of 72ºF. The 25th percentile was 64ºF, while the 75th percentile was 68.75ºF. The standard deviation was 4.55.

18Z dew points had a range from a low of 56ºF to a high of 73ºF. The 25th percentile was 60.75ºF, while the 75th percentile was 68ºF. The standard deviation was 4.70.

Finally, 21Z dew points had a range from a low of 59ºF to a high of 72ºF. The 25th percentile was 63ºF, while the 75th percentile was 69ºF. The standard deviation was 4.04.

 

 

Figure 88. Same as in fig. 86, but for EH type. 

 

For the HS type (fig. 89), the median dew point increases from 64ºF at 12Z to 66ºF at 15Z then decreases to 63ºF by 18Z and then increases to 64ºF by 21Z.

12Z dew points had a range from a low of 57ºF to a high of 73. The 25th percentile was at 61ºF while the 75th percentile was at 68ºF. The standard deviation was 5.13.

15Z dew points had a range from a low of 55ºF to a high of 75. The 25th percentile was 62ºF, while the 75th percentile was 68.75ºF. The standard deviation was 5.49.

18Z dew points had a range from a low of 59ºF to a high of 73. The 25th percentile was 61.75ºF, while the 75th percentile was 69ºF. The standard deviation was 4.75.

Finally, 21Z dew points had a range from a low of 55ºF to a high of 77. The 25th percentile was 61.5ºF, while the 75th percentile was 67.75ºF The standard deviation was 6.39.

 

 

Figure 89. Same as in fig. 86, but for HS type.

 

It is interesting to note that median temperatures increase from 18Z to 21Z for PH types, while they decrease during this time frame for BH types. Perhaps the northwesterly flow in PH allows for more downsloping and thus higher temperatures. Also of note is that BH has the highest temperatures out of all the synoptic types at 12Z, perhaps this is due to the more humid nature of the synoptic type (a more humid air mass will prevent overnight temperatures from dropping too much due to the high specific heat of water), as is noted with the highest median dew point in table 14 (3-day Yarnal dew points) and it had the highest hourly median dew points out of all the Yarnal synoptic types.

There is little change overall from hour to hour in regards to dew point in the synoptic types, however, the least change occurs in the BH type, where the median dew point remains at 68 from 12Z through 18Z.

The PH median dew points show an increase through 15Z, then decreases through 21Z. This again may show how downsloping plays a part in these types of heat waves. Downsloping causes compressional heating and drying. Already previously noted was the warming through 21Z, and the decreasing dew points after 15Z is evidence of drying.

BH has both the highest and lowest observed dew points. This could be due to the fact that a transition from one synoptic type to another is usually observed. For instance, PH starts off, then transitions to EH, and finally BH. The lower dew points may actually be due to a quick transition from the dry PH overnight which then quickly becomes BH by 12Z (the time the maps were investigated to determine what synoptic type was affecting NYC that day).

EH had the lowest standard deviation in the 3-day heat waves among all synoptic types (table 14), however, it had some of the highest standard deviations when looked at hourly. In other words, EH dew points didn’t vary much as compared to other synoptic types, but when looked at hourly, it had some of the highest variation compared to the other synoptic types.

When looking at both high and low temperatures, HS seems to be the warmest synoptic type when looking at 3-day heat waves. It has the 2nd highest maximum temperature (tied with BH at 102ºF [table 9], the highest median temperature for the high and low temperatures, it has the highest temperature at the 25th percentile for both high and low temperatures and is tied for the highest temperature at the 75th percentile for high temperature and has the highest temperature at the 75th percentile for the low temperatures (tables 9 and 10). However, as previously noted, HS does not occur frequently.

In general, there isn’t much range between temperatures for the synoptic types for 3-day heat waves. Maximum high temperatures only range from 99ºF-102ºF and maximum low temperatures only range from 82ºF-84ºF. Minimum low temperatures only range from 58ºF-62ºF. Median temperatures for highs range from 92ºF-93ºF and from 72ºF-75ºF for lows. The 25th percentile ranges from 91ºF-92ºF for highs and 70ºF-72ºF for lows. Finally, the 75th percentile ranges from 94ºF-95ºF for highs and from 74ºF-77ºF for lows.

CONCLUSIONS

In general, high temperatures increase from the start of a heat wave, to somewhere in the middle, then decrease.  For 3-day heat waves, high temperatures increase from day 1 to day 2, then decrease from days 2 to 3. For 4-day heat waves, high temperatures increase from days 1 to 3, then decrease through day 4. For 5-day heat waves, high temperatures increase from days 1 to 3, then decrease through day 5. Low temperatures follow the same pattern.

Generally, longer heat waves mean higher temperatures. Low temperatures do not necessarily follow this pattern.

The frequency of heat waves may be another metric to gage climate change. In NYC, prior to the 1940s, there were few higher order (6-day or greater) heat waves. Beginning in the 1940s and onward, there were multiple higher order heat waves in every decade. Also, from the 1930s through the 1990s, the frequency of 3-day heat waves occur at least as often as the peak in the 3-day heat waves from the 1880s-1920s (peaking at 9 events in the 1890s). There were only 7 3-day heat wave events in the 2000s, however, this is misleading because there is a peak of 6 4-day heat waves in this decade (tied with the 1990s), plus there were 2 6-day heat waves and an 8- and a 9-day heat wave event. There may have been a peak of 4-day heat waves in the 1940s, but 4-day heat waves have been on the rise since the 1950s, with more higher order events occurring.

The earliest heat waves can occur in NYC is April and the latest is in September. The occurrence of heat waves increases from April through July (the peak), then they decrease through September.

850 hPa temperatures generally range between 15°C to 20°C during heat wave events. 850 hPa temperatures also generally follow the pattern of high temperatures, increasing through the middle of the heat wave then decreasing (when looking at 8 pm temperatures, slight variations are seen in 3-day heat waves for 8 am temperatures).

Looking at average maximum hourly temperatures and the dew points at those hours, 4-day heat waves have the highest maximum hourly temperatures and dew points and thus highest average HI, with the greatest difference between average actual maximum hourly temperature and average HI (4.1°F). This may contradict the previous statement that 5-day heat waves have higher maximum temperatures than 3- and 4-day heat waves, but there are two things to consider: 1) these are maximum hourly temperatures taken every 3 hours (hourly temperatures are not available for the entire data set, so every 3 hours was used). They are likely not capturing the high temperature for the days. 2) There were only 8 events to consider for 5-day events (40 temperatures) as opposed to 18 events (72 temperatures) for 4-day and 31 events (93 temperatures) for 3-day events. Thus, there may not be enough 5-day events to accurately compare to 3- and 4-day events.

A daytime diurnal curve can be developed for 3-, 4- and 5-day heat wave events. Temperatures increase from 12Z through 18Z, then decrease slightly by 21Z. The largest increase is seen from 12Z-15Z (by 8°F-9F). Thereafter, temperatures increase 5°F-6°F from 15Z to 18Z. Thereafter, temperatures drop slightly (up to 2°F). Combine this with the fact that dew points change very little throughout the daytime hours, there is an overall decrease in RH throughout the day. While a decrease in RH may seem like the air mass is getting drier and therefore the HI should not be as high, the temperature rise counteract this feature. Meaning that because RH is decreasing, it does not necessarily mean the HI is decreasing, since temperatures are increasing more, allowing the HI to increase despite the lowering of RH.

A majority of maximum hourly temperatures occur between 17Z and 19z, with 20Z-21Z coming in 2^nd. A small percentage of maximum hourly temperatures occur very early in the day, between 15Z-16Z. Increased cloud cover, convection or precipitation in general, and passage of a cold front all can lead to cutting off diabatic heating early in the day.

Most dew points at the time of max hourly temperature fall between 65⁰F and 69⁰F, though 70⁰F to 74⁰F is most common for 4-day heat waves, while there is a secondary peak of lower dew points for 5-day heat waves between 54⁰F and 59⁰F.  This, along with other data shown, proves that 4-day heat waves are generally associated with higher dew points (and thus higher heat indices) while 5-day heat waves are associated with lower dew points (and thus lower heat indices). However, based on research from Matte et al. XXXX (date of article), prolonged exposure to heat waves for longer periods of time that are not as hot as heat waves that are hotter but shorter in duration may be equally, if not more dangerous.

Comparisons of temporal evolution of composite MSLP from 3-, 4-, and 5-day heat waves show that a Bermuda type high pressure system is in place over the northern Atlantic Ocean for all 3 heat wave types, with some differences in the average central location of the high. During the 3-day heat waves, this high pressure intensifies on the second day, then weakens by the third day. 5-day weaken on the 2nd day, then remains at the same strength through the 5th day. For 4-day heat waves, the high remains nearly the same strength through all 4 days on average. All three heat wave types exhibit a thermal trough that develops over the NEUS, with slight differences in spatial and temporal evolution. MSLP anomalies show a range from 0 to +2 mb over NYC at the beginning of 3- and 4-day heat waves, then trend to -1 to -3 mb near the end. 5-day heat waves begin with up to a +3 mb to +4 mb anomaly, then trend to 0 through the period, with slightly negative anomalies located just north of NYC by day 5.

Comparisons of temporal evolution of composite 500 mb height shows that the NEUS lies between an exiting trough and entering ridge. The ridge (over the Great Lakes region) typically deamplifies (weakens) as it moves east during the heat wave events, becoming zonal near the end of the events. A closed off high is also evident over the southern US, and usually retrogrades west through the events. 500 mb height anomalies show some differences between the 3 different types of heat waves, but generally, positive anomalies exist over the eastern half of the US, with a core of +80 mb or greater height anomalies over or just north of the Great Lakes region. This core remains steady then weakens for 3- and 5-day heat waves, but strengthens on the 2nd day of 4-day heat waves, then weakens through the end of the heat wave.

At 850 mb, temperatures over NYC start off at 14⁰C to 16⁰C and generally warm to 18⁰C to 20⁰C through the end of all heat waves. 850 mb temperature anomalies start off +3⁰C to +5⁰C over the Great Lakes region. This anomaly stays constant in magnitude for 3-day heat waves, but becomes smaller in areal coverage as it moves east through the 3 day period. For 4-day heat waves, the anomaly increases to +5⁰C to +6⁰C through about half way through the heat wave, before decreasing somewhat at the end. For 5-day heat waves, the magnitude of the anomaly remains constant (>+4⁰C) through day 3, then weakens thereafter to between +2⁰C and +3⁰C by day 5. 

Surface winds for 3- and 4-day heat waves generally show no change in direction. Speeds range from 8-9 mph for 3-day heat waves and 7-8 mph for 4-day heat waves. 5-day heat waves show an increase in wind speeds from 5.0 to 5.6 mph on day 1 to 8.4 to 8.9 mph on day 4, then decrease to 7.8 to 8.4 on day 5. 5-day heat waves do show some minor variations in wind direction, from west on day 1 to southwest over the rest of the event. In all heat waves, the strongest winds are offshore.

Recognizing the above patterns in the composites is key to forecasting heat wave events in Central Park. Anomaly composites presented here can be used with Pennsylvania State University-National Weather Service's Weather and Research Page and analyzing the anomaly maps for the same elements presented here. If models show the same or similar patterns presented in this research, a heat wave is likely to occur. The web site is: http://cms.met.psu.edu/sref/

The most common high pressure synoptic type to affect NYC during any heat wave is a BH type (which affects anywhere from 33%-58% of all events depending on which type of heat wave). This is a Bermuda high type high pressure system that introduces a warm and humid air mass into the region. The next most common type to affect NYC is EH for 3-day heat waves, and PH for 4- and 5-day heat waves. The center of high pressure is over the area in EH, and subsidence occurring underneath this high allows for compressional heating. With PH, downsloping winds off the higher terrain from the northwest, especially after the passage of a cold front in the summer, when winds are typically stronger, allows for compressional heating and drying at the coastal plain (where NYC is located) as well. HS is the next most common type. The westerly component here also allows for some compressional heating.

WORKS CITED

Burrows, A.T.: Hot waves: conditions which produce them, and their effect on agriculture. Yearbook of the United States department of agriculture.  Washington D.C.:  U.S.  G.P.O., 1901. 325-336.

Cubasch, U. et al, 2013: Introduction. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Kalnay, E. and Coauthors, 1996: The NCEP/NCAR Reanalysis 40-year Project. Bull. Amer. Meteor. Soc., 77, 437-471.

L. P. Rothfusz, “The Heat Index Equation (or, More Than You ever Wanted to Know about Heat Index),” Tech. Attachment, SR/SSD 90-23, NWS S. Reg. Headquarters, Forth Worth, TX, 1990. http://www.srh.noaa.gov/images/ffc/pdf/ta_htindx.PDF

Metzger KB, Ito K, Matte TD (2010) Summer heat and mortality in New York City: how hot is too hot? Environ Health Perspect 118: 80–86. doi: 10.1289/ehp.0900906 [PMC free article] [PubMed].

National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis (data available from 1948-2017) and the North American Regional Reanalysis (NARR) (data available from 1979-2017).

National Weather Service Glossary of Meteorology, published online 25 June, 2009 1:01 PM from http://w1.weather.gov/glossary/ retrieved June 9, 2012.

NOAA Central Library, https://library.noaa.gov/Collections/Digital-Collections/US-Daily-Weather-Maps, retrieved April 4, 2012.

NOAA National Centers for Environmental Information, https://www.ncei.noaa.gov/ retrieved August 2-4 2011. 

Oolman, Larry, University of Wyoming, College of Engineering. http://weather.uwyo.edu/upperair/sounding.html retrieved September 2nd, 2013.

Pennsylvania State University-National Weather Service's Weather and Research Site, published online July 19, 2010,  retrieved April 15, 2018 from http://cms.met.psu.edu/sref/.

Pollina, J. B. et al, 2013: Climatology and meteorological evolution of major wildfire events over the northeast united states, Weather and Forecasting, 28 (1), 175-193.

Robinson , P.J., 2001: On the definition of a heat wave. Journal of applied Meteorology, 40, 762-773.

Solecki, W.D. et al, 2005: Mitigation of the heat island effect in urban New Jersey. Glob Environ Change Part B: Environ Hazards 6 (1), 39–49.

United States Census Bureau, American Housing Survey (AHS), https://www.census.gov/programs-surveys/ahs/data.All.html, retrieved April 17 2012.

United States Environmental Protection Agency, 2012, https://www.epa.gov/heat-islands, retrieved online July 6, 2012.

WPC’s Surface Analysis Archive, http://www.wpc.ncep.noaa.gov/archives/web_pages/sfc/sfc_archive.php, retrieved April 4, 2012.

Wunderground, https://www.wunderground.com/history/airport/KNYC retrieved May 2, 2012.

Yarnal, B. 1993:  Synoptic climatology in environmental analysis:  a primer.  Belhaven Press, London.