National Weather Service United States Department of Commerce


Fire Weather Climatology of the Northeast U.S.



QUICK LINKS TO SECTIONS:

 
 

QUICK LINKS TO FIGURES:
 
 

DATA AND METHODS

a. Fire weather climatology over the Northeast U.S.

The Northeast U.S. (NEUS, includes PA, NJ, NY, CT, RI, MA, VT, NH, and ME) is very diverse in regards to population density and forestation. Much of the NEUS is made up of sparsely populated forested regions (central PA, southern NJ, southern and northern NY, western MA, and much of VT, NH, and ME). It also has highly populated areas that have no forests or very little in the way of forestation (specifically the megalopolis from Philadelphia through NYC and western and central Long Island, into western CT, and Boston). The areas just outside the megalopolis (speckled blue, hatched maroon, and red colors in the picture below (fig. 1 from the United States Forest Service, 2010) show areas of highly populated forested areas. Wildfires in these areas would have the greatest impact to life and property.

 
 
 
 
Figure 1

 

The fire weather climatology over the NEUS was constructed using actual wildfire days, which consists of 155 major (>100 acres burned) wildfires across the NEUS from January 1999 to December 2009 (fig. 2). The wildfire data were obtained through the Northeast Interagency Coordination Center (NICC) and the Pennsylvania Bureau of Forestry. Since topography, terrain, population, and land surface characteristics are key factors for wildfires, the NEUS was divided into two subregions for this study. Region 1 encompasses much of the higher elevations (typically greater than 1000 ft. in elevation) of the NEUS (fig. 2), while the lee of the Appalachians and most of the coastal plain make up region 2, which includes many of the more urbanized areas. The thick black line in figure 2 shows the line of delineation between the 2 regions. When identifying the days for the actual wildfire, multiple wildfires that occurred for a region on a particular day were only counted once; therefore, the 155 wildfires occurred on 42 separate days in region 1 and 73 in region 2. About 61% (96 out of 155) of all actual wildfires occur along the coastal plain in region 2, while 39% occur in region 1. Thus, region 2 has ~0.37 fires per 1000 square kilometer for the 11 year period, while region 1 has ~0.20 fires per 1000 square km. Wildfires are clustered in MA, portions of southwest CT westward to the Lower Hudson Valley in NY, the Pine Barrens of southern NJ in region 2, and in central and northeastern PA in region 1.  These areas in region 2 coincide with heavily populated forested regions of the NEUS (figure 1). The shading in figure 2 is elevation height, with darker shadings denoting higher elevation.

Figure 2

Back to Top
The Yarnal (1993) synoptic classification system was used to determine the large scale pressure patterns at the surface associated with the actual wildfire days. Yarnal identified eight different types of surface pressure patterns over the NEUS. Figures 3 and 4 show examples of five Yarnal synoptic types used for this study. The “pre-high” synoptic type occurs when surface high pressure builds southeastward over the NEUS, 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”, 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, which allows for southwesterly flow over the NEUS region. With “high to the south”, 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.  Other synoptic patterns from the Yarnal classification system which aren’t shown are an “elongated low” which is an elongated area of low pressure, covering a relatively large area, sometimes with multiple centers.  “Cyclonic with rain” is when the NEUS is under cyclonic flow with rain.  “Cold front” is when a cold front moves through.  Finally, “trough” (not part of the Yarnal classification system) is when there is an elongated region of relatively low atmospheric pressure at the surface that is not associated with a cold front.
 

Figures 3 and 4

The National Centers for Environmental Prediction (NCEP) surface synoptic weather maps for 1200 UTC available from the University of Washington (https://www.atmos.washington.edu/data/vmaproom/varchive.cgi) were manually inspected to determine the Yarnal classification of each actual wildfire case in the dataset. If there were missing images from the University of Washington web site, then the analyses from other web sites, such as the Storm Prediction Center (http://www.spc.noaa.gov/obswx/maps/), Weather Prediction Center (http://www.wpc.ncep.noaa.gov/html/sfc_archive.shtml), and Plymouth State (https://vortex.plymouth.edu/u-make.html) were used.  The North American Regional Reanalysis (NARR; Mesinger et al. 2006) was used to composite the large-scale flow evolution and other meteorological variables for the actual wildfire events. Spatial composites were created for regions 1 and 2 separately using the actual fires dates.  Since wildfire start times for many fires were not available, daily composites were created using 3 hourly NARR files from 0000 UTC (Universal Time Code) to 2100 UTC on the date the fire was reported for Mean Sea Level Pressure (MSLP) and 500 hPa heights.  A daily composite of RH and lower level winds would give a false sense of what these conditions were really like during the start of the fire, since these values vary greatly diurnally.  Therefore, the 2100 UTC NARR files were used to composite 2-m RH and 925 hPa winds on the date the fire was reported, since this is typically the warmest and driest time of day.  For region 1, a composite average using all fire dates for all the states tended to smooth the different fire weather patterns from west to east, so PA was not included in the region 1 NARR composites.

d. Trajectory Analysis

The Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model (Draxler and Rolph, 2003) was used to determine the origin of the air near the top of the boundary layer (~850hPa). HYSPLIT computes three-dimensional parcel trajectories using the NARR reanalysis available every 6 h (Mesinger et al. 2006). Backward trajectories at the surface were calculated for 48 h prior to 2100 UTC on the date of each actual wildfire, since this is typically the warmest and driest time of day.  Average height, temperature, and RH were calculated for the common synoptic patterns to affect the NEUS during major wildfires.
 

Outside of the heavily populated, nonforested metro areas of the NEUS such as New York City, Northeast New Jersey, eastern MA, and southeast PA, the coastal plain of the NEUS is a heavily populated forested region, which can be sensitive to the effects of wildfires.  Small wildfires (compared to the Western U.S.) can have a large impact.  Given the impact that wildfires can have on life and property, especially across the highly populated region of the NEUS, it is important to better understand the ambient conditions that increase the likelihood of wildfires in this area. A comprehensive fire weather climatology will help forecasters recognize the features that are associated with the increased risk of NEUS wildfires. In particular, this research will address the following questions:  

 

  • Where are wildfires favored across the NEUS and how do they vary monthly and interannually?

  •   Does the wildfire climatology change for wildfire threat days as compared to actual wildfire days

  • What are the most common synoptic weather patterns associated with wildfires over the NEUS, and how do these weather patterns evolve?

  • What is the origin of the dry air that enters the planetary boundary layer for these events?

 RESULTS

Figure 5 shows the ignition source of the wildfire and the acres burned by those wildfires in the NEUS, which was compiled from data received from the Northeast Interagency Coordination Center (NICC) and the Pennsylvania Bureau of Forestry.  Basically, these are the causes of major wildfires from Jan 1999-Dec. 2009.  Light blue is acres burned by humans.  Dark blue is number of fires caused by humans.  Yellow is acres burned by other than human meansâ�¦e.g. nature.  Green is number of fires caused by other than human meansAlthough 23% of all acres burned by major wildfires from 1999-2009 were caused by aircraft, this total was due to one event (15,550 acres) that occurred on 15 May 2007 in the southern Pine Barrens near Barnegat, NJ.  Humans are the main cause for wildfires, with approximately 71% of all events in the NEUS caused by humans. Lightning represents only a small (~0.8% of all acres burned, or ~1.9% of all events) fraction, since the NEUS does not typically experience dry thunderstorms, which are responsible for many of the wildfires that occur in the western US (Rorig and Ferguson 2002).

Figure 5

Back to Top

Figure 6 shows the annual percentage of actual wildfire events that occurred from 1999-2009 for regions 1 and 2. The wildfire frequency in region 1 peaks in 2006, with ~24% (10 fires) of the 42 fires in the climatology occurring that year. All 10 wildfires in region 1 in 2006 occurred in the spring (March, April, and May), and 6 (60%) of the wildfires occurred during May alone (not shown). 2002 and 2005 were also active years for wildfires in region 1.  There were 7 wildfires in 2002, with 4 of the 7 wildfires (~57%) occurring during the spring (3 of the 4 in April), while the other 3 (~43%) occurred in the  summer (June, July, and August). In 2005, there were 8 wildfires to affect region 1, with 6 wildfires during April and May.  Wildfire frequency in region 2 peaks in 1999, with 22 out of 73 (~30%) occurring in this year, and 10 of these 22 wildfires (45%) during the spring (7 in April).

Figure 6

Back to Top

Investigating the monthly distribution of wildfire events will determine what time of year wildfires are most common for the NEUS (figure 7). There is a peak in actual wildfire events in April in both regions, with ~45% and ~34% of the fires occurring in this month in regions 1 and 2, respectively. April and May comprise ~76% and ~53% of the fires occurring in regions 1 and 2, respectively. During early to mid spring, the vegetation across the NEUS has not fully greened up, thus there are still dead leaves and twigs on the forest floor.  This dead vegetation cannot retain water as well as living plants, and even after a period of significant rainfall the dead vegetation dries out rapidly.  Li et al. (2010) concluded that from 2001 through 2007 the green up onset date for latitudes between 40°N and 45°N (southern sections of the NEUS) is between April 10th and 20th, meaning that much of the NEUS is still not green after about the first 2 weeks in April. Additionally, solar radiation and temperatures increase, which allows the fuels to dry out quicker (Parr et al. 2005, Kassomenos 2010).

In contrast, during the summer (June, July, and August) only ~10% of the actual wildfires occur in region 1 and ~22% in region 2 (figure 7). During summer the live vegetation typically holds abundant moisture, which prevents the fuels from igniting easily, and the relatively humid conditions in summer also are less favorable for wildfires. During the winter (December, January, and February) there is little or no fire activity (~0% for region 1 and ~3% for region 2), since the ground is cool, damp, and snow covered on average.

 

Figure 7

Back to Top

Figure 8 shows the climatology of synoptic flow patterns for the actual wildfire days using the Yarnal classification scheme. The pre-high (PH), back of high (BH), and extended high (EH) types together account for ~78% of the wildfire events in region 1 and ~77% in region 2.  The EH is the most common synoptic type in region 1 (~45% of all events), while the PH type is most common in region 2 (~30% of all cases). The combination of cold fronts (CF), elongated low (EL), high to the south (HS), surface trough and high to the north (HN) together account for ~21% and 23% of the fire events for regions 1 and 2, respectively. A CF is often associated with precipitation, so it has a relatively low percentage (~10% of all cases in region 1 and 14% region 2). The EL type is also often associated with precipitation, and thus there were only a few wildfires in both regions associated with this synoptic type. Meanwhile, a HN (0% in both regions 1 and 2) allows an easterly flow to develop, which advects cool and moist marine air from the Atlantic Ocean towards the NEUS during the peak months of wildfire activity.

Figure 8

Back to Top

Figure 9 shows the monthly percentage for each synoptic type for the actual fire days in region 1. The EH type is the most common synoptic type to occur in April (peak of wildfire season) at 21% (occurs 12% more than the next most common synoptic type, CF, and 14% more than the other synoptic types that occur, PH and BH) and occurred 19% of the time in May.  Major wildfires with CFs (~10%) only occurred in April in region 1, while trough events (~5%) only occurred in May. EH and BH are the dominant synoptic types associated with major wildfires in July and August in region 1 (~2% for both synoptic types in July, and 5% for BH and 0% for EH in August). 

Figure 9

Back to Top

For region 2 (figure 10), although the PH type is the most common pattern overall (~27%), it is not the most common type in any individual month. Rather, during the peak of fire season in April, the PH is similar to BH (~10%), while the EH is the dominant type in May (~8%). Wildfire activity associated with CFs is mostly in April, but it also occurs in the months of May, June, July, and October.

Figure 10

Back to Top

Spatial composites were created using the North American Regional Reanalysis (NARR) for actual wildfire days in regions 1 and 2 separately.  At 48 h prior (t-48) to a fire event in region 1 (figure 11a), there is a surface high pressure (lines are isotherms) over the eastern Great Lakes, while there is a weak surface low just east of the mid-Atlantic states.  The 925 hPa wind directions (barbs) range from westerly and northwesterly over southern and western sections of the NEUS to north-northwesterly and northerly over northern and eastern sections of the NEUS.  Wind speeds are 5 m s-1 and below (half barb=5 m s-1, full barb=10 m s-1).  Potential temperatures (shaded) range from between 288K–293 K in extreme southern PA to 278K–283 K in northern NEUS, with a majority of the NEUS between 283K–288K.  On the day of the event (t=0) in region 1 (figure 11b), the center of the surface high moves south, located just south of the NEUS, over West Virginia and Virginia, and is similar to the EH pattern in the Yarnal classification, while the surface low off the mid-Atlantic states has moved slightly east. The 925 hPa wind speeds over the NEUS are generally northwesterly, with some westerly and southwesterly winds in southwest PA, and northerly winds along southern coastal sections.  Wind speeds are still 5 m s-1 or less.  Warmer potential temperatures have advected northeast slightly and range between 283K-293K across the NEUS.

 

 

Figure 11

Back to Top

There is a broad ridge at 500 hPa (lines are 500 hPa heights in decameters) at t-48 (figure 12a), from the western Great Lakes south-southwestward to Texas.  Winds at this level (barbs, same set up as previous figure) are northwesterly at about 10 m s-1.   The NEUS lies between two jet streaks at 300 hPa (shaded), one well off the NEUS coast oriented southwest to northeast, situated just east of the 500 hPa trough axis that is over the western Atlantic with a jet core of 25 m s-1 to 27 m s-1.  The other jet streak is north of the western Great Lakes in southern Canada oriented west to east with a core of 23 m s-1 to 25  m s-1.  Winds speeds over the NEUS are between 15 m s-1 and 17  m s-1.   At t=0 (fig 12b), the 500-hPa ridge axis has moved to the central and eastern Great Lakes and has amplified, while the upper level trough is still situated ~750 km off the U.S. East Coast and has also amplified.  Winds at this level remain northwesterly at about 10 m s-1.  The jet off the NEUS coast has pushed farther east, while the jet over southern Canada has moved southeast and is situated north of the Central Great Lakes region over southern Hudson Bay. The core has strengthened to between than 27 and 29 m s-1.  The NEUS lies in a somewhat favorable area for subsidence to occur.  It is a well known fact that within jet circulations, subsidence occurs over the right exit regions of upper level jets.  The NEUS lies near this position, especially northern portions of the NEUS.

 

Figure 12

Back to Top

Subsidence is observed over the entire NEUS in a plot of average 600 hPa to 300 hPa vertical velocity (shaded, Pa/s) for region 1 at t0, with some enhanced sinking occurring over central NY and northern VT and NH (figure 13). Contours are 300 hPa wind speeds.  As discussed in the previous figures, the NEUS lies on near the right exit region of a jet streak that is located just north of the Central Great Lakes region.

Figure 13

Back to Top

Figure 14a shows the composite 2-m RH (shaded), 925 hPa wind (barbs, same set up as previous figures), and 2-m temperature (dashed line) for region 1 at t-48 and t=0 (figure 14b), the day of the event.  The lowest RHs (45-55%) are from the majority of PA and stretches east and northeast into all of NJ, New York City, the Lower Hudson Valley, western, northern, and northwest CT, northern RI, western MA, and southern NH at t-48h.  2-m temperatures generally range from 11°C – 13°C in ME, to 18°C in extreme southern and southwestern PA. There is a temperature gradient along the immediate coast where the relatively cool waters meet the warm inland temperatures, and temperatures here are as low as 9°C.  At t=0 in region 1 (figure 14b), the RHs decrease to less than 45% across extreme southern NJ and extreme southern PA.  A larger area of 45%-50% RH values exists over central and eastern PA and most of NJ, which extends into southern NY and the lower Hudson Valley.  The rest of the NEUS has RH values of between 55%-65%.  This composite analysis suggests that these fire events can occur even though the RH values are greater and wind speeds are less than the criteria for a RFW (30% RH and 11.2 m s-1 wind speed). However, comparing NARR and the actual surface observations for a few cases, the NARR was too moist by 15-30% (not shown), therefore illustrating the difficulty in obtaining an accurate low-level moisture analysis for these relatively dry fire events.  The 2-m temperatures have generally increased 2°C-3°C at t=0, with the warmest air over PA where temperatures range from 18°C-21°C.  Northern sections of the NEUS have warmed to generally between 13°C-15°C, with the immediate coastal sections as low as 11°C.

Figure 14

Back to Top

Figure 15a shows a cross section of potential temperature (solid lines), winds (barbs, same format as previously), and vertical velocity (shaded, Pa/s) along latitude 42.2N at t-48 from longitudes 70.78W to 79.5W.  From 73.0W to 70.78W is region 2, and from 73.0W to 79.5W is region 1 and is denoted by the vertical line.  There is some instability in the lower levels between 74.5W and 78.5W in region 1 as the 280K isentrope slopes.  There is upward motion in the lower levels from just above the surface to around 800 hPa, with maximum upward motion seen at around 75.0W, with values between -0.15 Pa s-1 to -0.10 Pa s-1.  This upward motion in the lower levels of region 1, and downward motion in region 2 is associated with upsloping and downsloping winds, as the lower levels have enhanced vertical velocities (negative for region 1 and positive for region 2), while at the same time a westerly or northwesterly cross barrier flow exists.  There is downward motion associated with synoptic subsidence from 800 hPa to about 300 hPa over region 1. The west to northwest winds at the lower levels veer to the northwest for western portions of region 1 indicating an area of warm advection for extreme western portions of region 1 from 78.0W to 79.0W.  At t=0 (fig. 3.12b), the area of instability in the lower levels moves east, between 73.0W and 76.0W with a sloping 285K isentrope, while the upsloping continues in region 1 and downsloping occurs in region 2 in association with negative and positive vertical velocities respectively in the lower levels, and synoptic scale subsidence occurs from about 900 hPa and above. 

Figure 15

Back to Top

For region 2 at t-48 h (Fig. 16a, same format as figure 11), there is an area of surface high pressure near western North Carolina and eastern Tennessee, and a surface trough just off the east coast, with the trough axis extending from 500 km to 650 km east of the coast of Georgia, northeast to between 800 km and 900 km east of Virginia. The 925 hPa winds are generally northwesterly at about 5 m s-1 or less, and potential temperatures range between 283K-288K over southern NEUS in PA to 273K-278K over extreme northern NEUS in northern ME.  A majority of the NEUS is between 278K-283K.  At t=0 (Fig. 16b), the surface high pressure has remained nearly stationary with a slight increase to 1022 hPa. As a result, a PH type pattern is established across the NEUS, with westerly or west-northwesterly 925 hPa flow across the NEUS.  Wind speeds are still at 5 m s-1 or less with nearly the same potential temperature range as t-48.

 

Figure 16

Back to Top

In the upper levels, at t-48 (Fig 17a, same format as figure 12), there is a broad 500-hPa ridge over the central portions of the U.S., while an upper-level trough is located ~750 km east of the East coast. 500 hPa winds are northwesterly at 20 m s-1.  At 300 hPa, there is a jet oriented east to west from south central Canada and the Northern Plains to southeastern Canada, with wind speeds of 21 m s-1 to 23 m s-1.  By t=0 (Fig. 17b), the upper level pattern is similar to 48 h earlier, with wind speeds of 15 m s-1 at 500 hPa.  At 300 hPa, there is a jet streak over southeastern Canada oriented northwest to southeast, with the a jet core of between 25 m s-1and 27 m s-1 just north of the eastern and central Great Lakes region and south of the Hudson Bay.

Figure 17

Back to Top

 

As with region 1, subsidence is observed over much of the NEUS in plots of 600 hPa to 300 hPa averaged vertical velocity for region 2 at t=0 (figure 18, same format as figure 13), with some enhancements over northern VT, NH, and ME, with no real enhancements noted for portions of region 2.  It cannot be made clear from this plot whether the jet is enhancing the subsidence over this region.

Figure 18

Back to Top

The lowest surface RHs (45%-50%) occur over extreme southeastern PA and southern NJ, but a larger portion of low RH (50%-55%) exists over the coastal plain (Fig 19a, same format as figure 14), from southern NH to New Jersey at t-48. This region of lower RHs corresponds to the cluster of past wildfires (Fig. 2.1). The location of the lowest RH for t-48 (fig. 2) and t=0 (fig. 19b) is east of the Appalachian Mountains.

Figure 19

Back to Top

 

The westerly component favors downslope drying over the coastal plain, with relatively strong low level subsidence noted from the surface to about 700 hPa (values of +0.09 Pa/s to +0.12 Pa/s) in cross sectional composite at t-48 (fig. 20a, same format as figure 15). Synoptic scale subsidence exists above 700 hPa with lower values of positive vertical velocity (values of 0 Pa/s to +0.09 Pa/s).  Once again, lift is seen in association with cross barrier westerly flow leading to upsloping in region 1 with negative values of vertical velocity.  Much of the same is seen at t=0 (fig. 20b)

Figure 20

Back to Top

In order to gain an understanding of where the subsidence is coming from in a quasi-geostrophic sense, the Sutcliffe-Trenberth form of the quasi-geostrophic omega equation (Trenberth 1992) was used.  This equation states that anticyclonic vorticity advection by the thermal wind is associated with downward motion (term A in the Sutcliffe-Trenberth QG equation).  Figure 21 shows the 600 hPa to 300 hPa thermal wind and term A (vorticity term) composite for regions 1 (figure 21a) and 2 (figure 21b) for t=0.  Both regions show anticyclonic vorticity over the NEUS, and more being advected in by the northwesterly direction of the thermal wind crossing over the gradient of anticyclonic vorticity in a region to the to the northwest of the NEUS.  The values for region 1 are stronger as compared to region 2.

Figure 21

Back to Top

Figure 22 shows a box and whisker plot of the RH (fig. 22a) and wind speed (fig. 22b) for both regions 1 and 2 at t=0.  The white line is the areal mean.  The mean areal average for region 1 is ~60%, while for region 2, it is ~63%.  For wind speed, the areal average for region 1 is 5.7 m s-1, while for region 2 it is 6.3 m s-1.  The whiskers show the minimum and maximum areal average values observed on the wildfire dates for each region.  While both regions share the same minimum areal average value RH of ~43%, major wildfires in region 2 have occurred at much higher areal average RH values than region 1 (~89% and 74% respectively).  Areal average wind speed minimum and maximum values seem to be similar for regions 1 and 2, with minimum values of 0.23 m s-1 and 0.73 m s-1 and maximum values of 14.1 m s-1 and 15.42 m s-1 respectively.  The standard deviation for region 1 for RH is 8.5 and for region 2 is 9.5, so the maximum RH value of 89% is ~2.7 standard deviations from the mean, while the 74% maximum value in region 1 is 1.7 standard deviations away.  The minimum RH values are 2.1 standard deviations from the mean for region 2 and 2.0 standard deviations for region 1.  This data suggests that wildfires occur in region 2 at a larger range of RH.  The standard deviations for wind speed are much closer for both regions, with a standard deviation of 3.0 in region 1 and 2.8 in region 2, indicating that wind speed does not differ much between the 2 regions for wildfires.

Figure 22

Back to Top

The synoptic composites show that wildfires can occur for different surface pressure patterns over the NEUS.  High pressure is the result of subsidence, and subsidence creates warmer and drier conditions.  In order to determine the origin of the subsidence and dry air on the large scale, back trajectories starting later in the afternoon at 2100 Universal Time Code (UTC) (start times of fires are unavailable so 2100 UTC was used, since this is typically the warmest and driest time of the day) were run to 48 hours prior, starting at 1500 m above sea level (this height was chosen because it has been shown that air from upper levels of the atmosphere descend to the midlevels [Charney 2009]).  Since subsidence is sinking air, i.e., downward vertical motion, the average height along 6 different trajectories (fig. 23) of the 3 major synoptic types to affect the NEUS during wildfire events (PH, EH, BH) were calculated. 

Figure 23

Back to Top

Figure 24 shows the average of all 28 PH heights along the six different trajectories.  Six trajectories were chosen to gain a good enough spread across the NEUS, while at the same time it is a low enough number of trajectories to easily see each trajectory.  Figure 23 shows the location of the start points of the trajectories.  #1 is southern NJ, #2 is central PA, #3 is the lower Hudson Valley in southern NY, #4 is in CT, #5 is in MA, and #6 is in VT.  These points were chosen to be at or near the cluster of major fires discussed earlier.  Because PA was the only state that had a cluster of large wildfire activity in region 1, VT was chosen randomly as another point for region 1, otherwise there would only be 1 point representing region 1.  On average, during a PH event, trajectories at 1500 m above ground level start between 2000 and 3500 meters above ground 48 hours prior (fig. 24), with 4 out of 6 trajectories starting between 3000 and 3500 meters.  This corresponds to about 500 meters to 2000 meters of descent. Trajectory #2 shows the most subsidence, starting out at  ~3350 m on average 48 hours prior, while trajectory #6 shows the least subsidence, only starting out at ~2100m.

 

 

Figure 24

Back to Top

Figure 25 shows the average of all 32 EH heights, with all the trajectories starting out 48 hours prior with heights ranging from 2250m. 

Trajectory #2 shows the most subsidence as it has the highest starting point (~2700m), while trajectory #6 has the lowest (~2350m).

Figure 25

Back to Top

Figure 26 shows the average of all 24 BH events.  The trajectories start out between 1750m-2250m 48 hours prior to the event.  All but 1 trajectory (#1) fall below 1500m starting 27 hours prior (1800 UTC on day 2, trajectory#6) and then rise to 1500m to the start of the trajectory on 2100 UTC on day 1.  This suggests that from 27h to 18h prior to a wildfire event in BH cases there is some rising motion occurring.  Out of all 3 synoptic types, PH seems to subside the most, while BH subsides the least, and even shows some lift just before the event.

Figure 26

Back to Top

Figure 27 shows the average RH along the 6 trajectories for the PH type.  RHs start out between 45%-60% 48h prior to the events. The largest decrease in RH occur with trajectories 1, 3, 4, and 5.  They start out between 50%-55% and end at ~35%, which is a 15%-20% drop in RH.  The largest drop off in RH seems to occur between 2100 UTC on day 3 to 0300 UTC on day 1, and levels off thereafter.  Trajectory 6 shows the least drop off in RH (~12% decrease). 

Figure 27

Back to Top

Figure 28 shows the average RH along the 6 trajectories for the EH type.  RHs start out between 38%-49% and end between 35% and 42%.  While PH shows a steady decrease throughout much of the individual trajectories (with the exception of some minor increases), EH shows a steady increase in RH beginning 1800 UTC on day 2 and ending on 1200 UTC on day 1 before dropping off again through 1800 UTC on day 1. 

Figure 28

Back to Top

As a whole, the BH type shows little change in RH from the beginning to the end of the trajectories (fig. 29).  Trajectories start out and end between 40%-50%.  Trajectories 1, 2, 3, and 6 show slight increases, while 4 and 5 show slight decreases.  Out of the 3 synoptic types, PH shows the most decrease in RH, while BH generally shows a slight increase.

Figure 29

Back to Top

Figure 30 shows the average temperature along the 6 trajectories for the PH type.  Trajectories start out between -17°C to -9°C 48h prior and end between -4°C and +4°C.  Five out of the six trajectories show and increase in temperature of between 15°C and 20°C, while one trajectory (#6) shows only ~6°C increase.

Figure 30

Back to Top

For the EH type (fig. 31), the temperature curves for all 6 trajectories are very similar.  Temperatures start out between -5°C and 0°C and increase to +5°C and +10°C, leading to a 10°C increase in temperature. 

Figure 31

Back to Top

The BH temperature curve (fig.32) for the 6 trajectories are also very similar.  Temperatures start out between +5°C and +10°C and end between +10°C and +15°C.  This corresponds to ~5°C increase.  Comparing the 3 synoptic types, PH shows the greatest increase in temperature, but the actual temperature values are the lowest.  BH shows the least temperature increase, but the actual temperature values are the highest.  Finally, EH is in the middle of the other two.

Figure 32

Back to Top

For the PH type, trajectory 6 is an outlier, showing the lowest decrease in height, the lowest decrease in RH, and the lowest increase in temperature.  Trajectory 6 is in region 1 in VT.  The other region 1 trajectory (#2 in PA) shows one of the highest decreases in height and temperature.  Meanwhile, the trajectories in region 2 (1, 3, 4 and 5) show similar decreases in altitude, RH, and increases in temperature compared to each other.  This may suggest that the  subsidence, decrease in RH, and increase in temperature may be more varied in region 1 than in region 2 during a PH type event, which is the predominant type to affect region 2.  In the other synoptic types, trajectory 6 is less of an outlier, indicating that the same values are less varied for these synoptic types, which includes EH (the predominant type to affect region 1).

The primary goal of this study was to develop an annual and monthly climatology for ACTUAL wildfires over the NEUS, but it does not include the large number of days with a wildfire THREAT.  A goal of the fire weather community is to alert the public when there is the potential for the ignition and spread of wildfires, which is defined as wildfire threat. Also, meteorological wildfire threat along the East Coast needs to be better understood (Charney et al. 2003; 2006; 2009). Therefore, in this section an abbreviated analysis of wildfire threat days will be shown, so it can be compared with the above actual wildfire days over the NEUS. The wildfire threat climatology was based on NFDRS ratings, since there were not enough RFWs issued to build a sufficient climatology. The climatology was created using days when 50% or greater of the NEUS had a NFDRS rating of “high” or greater, which was accessed from the Wildland Fire Assessment System (WFAS) database (www.wfas.net). There were 194 days from Jan 1999-Dec 2009 that met this criterion.  The WFAS archived NFDRS maps are only available for entire states so it was not possible to subdivide the events into the two regions shown in figure 2.

For the annual wildfire threat climatology (Fig. 33), there is a peak in 1999 and a secondary peak in 2006, which agrees well with the actual wildfire results (Fig. 6). Also, 2007 was one of the least active fire threat years (~2.1%), as well as the least active actual wildfire year for region 2 (~2.4%).  This is likely due to the abnormally wet and cool April 2007, which was the 3rd wettest and 25th coldest on record for the NEUS (Enloe 2010).

 

 

Figure 33

Back to Top

For the monthly climatology of wildfire threat (Fig. 34), there is an increase from January (~6%) to February (~13%). This result is slightly different than actual fires, where there is a significant increase of actual fire activity from March to April (Fig. 7). A possible explanation is that NFDRS does not output ratings when there is snow on the ground. This leads to a north-south gradient on the NFDRS wildfire threat map during February and March, where northern sections will have a low or moderate rating, while southern sections will have a high or greater rating.  The higher ratings for southern sections are enough to lead to an increasing number of February events when considering the whole NEUS.  Just over 30% of all days in February and March had this north-south boundary, and 100% of days with data available were associated with snow in northern NEUS. Also, since humans are the leading cause of wildfires in the NEUS, it stands to reason that although there is an elevated wildfire potential during the colder months of February and March, wildfires do not occur frequently due to a relative lack of human activity.

Figure 34

Back to Top

A simple conceptual model can be constructed with the synoptic composites that were developed to help aid forecasters.  For region 1 (fig. 35a), an extended high type is in place, with the center of the high located near the spine of the Appalachians, generally south of the NEUS, but extends into southern portions of the NEUS.  A weak surface low pressure system (solid lines are MSLP) is located well off the East coast in the western Atlantic.  At 500 hPa (dashed lines are 500 hPa heights), there is a ridge over the Great Lakes region, the axis of which extends into the Gulf States, with a trough of low pressure to the west of the surface low (generally located between the surface high and surface low).  Finally, at 300 hPa, there is a jet streak (shaded region) over southeastern Canada, just north of the Great Lakes oriented west to east.  For region 2 (fig. 35b), there is a surface high located at or just to the west of the spine of the Appalachians.  Since the center of the high is still to the southwest, high pressure is still building into the region, and this is classified as pre-high type.  A surface trough is well off the East coast in Western Atlantic.  The 500 hPa features are much flatter than in region 1, with a broad ridge over the Great Lakes and mid-West region.  At 300 hPa, there is a jet streak over southeastern Canada and the northern and eastern Great Lakes, extending into northern portions of the NEUS.  This jet is oriented northwest to southeast.

Figure 35

Back to Top

 

Conclusions

The goal of this study was to develop an annual and monthly climatology for the Northeast U.S. (NEUS) for wildfires.  The synoptic pattern in place during these events and their origin and evolution were determined, and the role of downsloping off the Appalachian mountains was addressed.  The annual and monthly climatology of wildfire events was summed over the NEUS for actual major wildfire events (> 100 acres). There was a peak in wildfires in 1999 for region 2 (coastal plain), and in 2006 for region 1 (interior Northeast).  These years were marked by abnormally warm and dry conditions. The wildfire season over the NEUS is April and May, with the peak occurring in April. This is likely due to the pre-green up period across the NEUS, increasing solar radiation, and the continental high pressure systems that move into the region during this time of year. There is a minimum in wildfire activity in the winter (DJF) due to the cool, damp conditions present during this time of year and the presence of snowpack.

Pre-high (PH), extended high (EH), and back of high (BH) are the 3 common types of high pressure systems to affect the NEUS during wildfire events.  PH establishes a flow that is favorable for downsloping, especially across region 2, where PH is most frequent.  EH is the most common type in region 1. This is usually a large area of high pressure that can take days to move across the region, and occurs during a time of year when blocking patterns are common for this section of the country.  The time it takes to move across the NEUS allows for persistent dry conditions, which permit the fuels to dry.  A southwesterly flow is normally established in a BH synoptic type.  The westerly component allows for some downslope and therefore dry conditions while the southerly component advects warm air into the NEUS.

Temporal evolution of synoptic plots show the origin of the high pressure systems are from the eastern Great Lakes region for region 1.  Over a 48 hour period, this high pressure moves to a position over the NEUS, classifying it as an EH type.  For region 2, a PH type is already in place over the NEUS, and is nearly stationary over the 48 hour period with some strengthening noted.  Little change is seen in RH values and 925 hPa wind speeds over region 1 when comparing t-48 and t=0.  Slight drying occurs in region 2.  Of particular interest with these variables is the fact that both variables lie outside the criteria for RFWs for the NEUS.  This begs the question “Do RFW criteria need to change for the NEUS?”  The NEUS is not subject to the same types of severe meteorological wildfire conditions as the western portions of the U.S., and therefore it is reasonable to assume that the NEUS may be more sensitive to the conditions that allow wildfires to grow because major wildfires occur at higher RH values and lower wind speeds for the NEUS as compared to the West.

HYSPLIT back trajectories show that the PH synoptic type undergoes the most subsidence, the largest decrease in RH, and the largest increase in temperature, while BH undergoes the least subsidence, least decrease (and in some cases an increase) in RH and least increase in temperature.  The EH type stands in the middle of these synoptic types.  The 850 hPa layer may be important when daytime heating allows for turbulent eddies to form and some of the drier air at this level is transported to the surface.  Also, as the air descends from 850hPa, temperatures will increase further, allowing fuels to dry out quicker.  This is supported by other studies previously mentioned.

The results from this study are useful in an operational forecast setting. For example, noting that a PH or EH type synoptic pattern will set up over the NEUS after a period of dry weather would be a signal to forecasters that conditions may be favorable for wildfires to occur.  In contrast, an EL, CF, HN, ST pattern will allow forecasters to realize that wildfires will be unlikely to occur. By describing the time of year when that NEUS experiences both wildfire threat and actual wildfires, a forecaster will be better prepared to look for conditions during the peak of wildfire season in April, while still noting that wildfires are possible but not as likely for other months.  As noted previously, RFW and FWW conditions vary from region to region, as well as their meanings.  For example, in some areas of the country, an RFW means that conditions are favorable to conduct a prescribed burn, while in others an RFW means that there are very unusual conditions and the fire management community must take the proper action to quickly respond to any reports of wildfires.  Therefore, understanding the frequency of actual wildfire days and wildfire threat is valuable information for the forecaster.

Criteria for RFWs are determined by the NWS along with representatives from the fire management community (Tim Morrin, NYC National Weather Service, 2011 personal communication).  However, these criteria are not necessarily set by a proper understanding of the frequency of high wildfire danger days and actual wildfire days.  This begs the question whether RFWs (and FWWs) capture an optimal percentage of high wildfire days or actual wildfire days. Also, as previously noted, the NEUS is more sensitive to wildfires than the western U.S. i.e., this study has shown that they can start at higher RH values and lower wind speeds than is set for the WFO OKX forecast area.  This study could be used to help reevaluate the criteria for RFWs and FWWs.

The information in this study may also be used to differentiate between unusual wildfire weather events and those that are more routine.  For example, if there is a high wildfire danger day expected in the month of January, the forecast can mention that this occurs once in ten years.

 

REFERENCES

 

Atger, F., 2003: Spatial and interannual variability of the reliability of ensemble-based probabilistic forecasts: Consequences for calibration.  Mon. Wea. Rev., 131, 1509–1523.

Avissar, R. and R. Pielke, 1989:  A parameterization of heterogeneous land surfaces for atmospheric numerical models and its impact on regional meteorology.  Mon. Wea. Rev., 117,  2113-2136.

Barriopedro, D. and Garcia-Herrera R.:  A climatology of northern hemisphere blocking.  Journal of Climate, 19, 1042-1063.

Bennett, M., 1995. A Lidar study of the limits to plume rise in a well-mixed boundary layer.

Atmospheric Environment, 29, 2275-2288.

Brockway, D. G., R. G. Gatewood, and R. B. Paris. 2002. Restoring fire as an ecological process in shortgrass prairie ecosystems: initial effects of prescribed burning during the dormant and growing seasons. Journal of Environmental Management 65, 135-152.

            Brotak EA, Reifsnyder WE (1977)  An investigation of the synoptic situations associated   

                        with major wildland fires.  Journal of Applied Meteorology 16(9), 867-870.

 

Buckley, C., 2008:  Forest Fire Scorches 3,000 Acres in Ulster Park.  The New York Times.  [Available online at https://www.nytimes.com/2008/04/21/nyregion/21minnewaska.html?ref=nyregion].

Buizza, R., 1997: Potential forecast skill of ensemble prediction and spread and skill distributions of the ECMWF ensemble prediction system. Mon. Wea. Rev., 125, 99–119.

Canadian Wildland Fire Information System, 2007:  Canadian wildland fire information system.  [Available online at https://cwfis.cfs.nrcan.gc.ca/en/index_e.php].

Charney, J. J., X. Bian, B. E. Potter, and W. E. Heilman, 2003:  The role of a stratospheric intrusion in the evolution of the Double Trouble State Park wildfire.  Proceedings, 5th Symposium on Fire and Forest Meteorology, Orlando, FL, 16-20 November, 2003.

Charney, J. J. and L. A. Fusina, 2006:  Employing numerical weather models to enhance fire weather and fire behavior predictions.  USDA Forest Service Proceedings RMRS-P-41.

Charney, J. J. and D. Keyser, 2009:  Mesoscale model simulation of the meteorological conditions during the 2 June 2002 Double Trouble State Park wildfire.  International Journal of Wildland Fire.  (19) 4, 427-448.

Chen, F., 2005:  Variability in global land surface energy budgets during 1987-1988 simulated by an off-line land surface model.  Climate Dynamics.  24(7-8),  667-684.

Chen, F. and R. Avissar, 1994a:  The impact of land-surface wetness on mesoscale heat fluxes.  J. Appl. Meteor.  33, 1324-1340.

__________and_____________, 1994b:  Impact of land-surface moisture variability on local shallow convective cumulus and precipitation in large-scale models.  J. Apple. Meteor.  33, 1382-1401.

__________ and J. Dudhia, 2001: Coupling an advanced land surface hydrology model with the Penn State–NCAR MM5 modeling system. Part I: Model implementation and sensitivity.  Mon. Wea. Rev., 129, 569–585.

Colle, B., 2001:  SUNY Stony Brook MM5 numerical forecast information.  [Available online at https://atmos.msrc.sunysb.edu/html/info.html].

Crimmins MA (2006) Synoptic climatology of extreme fire-weather conditions across the southwest United States. International Journal of Climatology 26(8), 1001–1016.

Desborough, C.E., 1999:  Surface energy balance complexity in GCM land surface models.  Climate Dynamics.  15(5), 398-403.

Dudhia, J., 1993: A nonhydrostatic version of the Penn State/NCAR Mesoscale Model: Validation tests and simulations of an Atlantic cyclone and cold front. Mon. Wea. Rev., 121, 1493–1513.

Draxler, R.R. and G.D. Rolph, 2003:  HYSPLIT model access via NOAA ARL READY website (https://ww.arl.noaa.gov/ready/hysplit4.html).  NOAA Air Resources Laboratory, Silver Spring, MD.

Eckel, F.A., and C. F. Mass, 2003: Towards an effective short-range ensemble forecast system. Proc. Workshop on Ensemble Forecasting, Val-Morin, QC, Canada. [Available online at https:// www.cdc.noaa.gov/people/tom.hamill/ef_workshop_2003.html.].

______________, and ___________, 2005: Aspects of effective short range ensemble forecasting. Wea. Forecasting, 20, 328–350.

Enloe, Jesse:  Plot time series national oceanic and atmospheric administration national climatic data center, retrieved March 3, 2011 [Available online at https://www.ncdc.noaa.gov/temp-and-precip/time-series/]

Enloe, Jesse: State of the climate national oceanic and atmospheric administration national climatic data center, retrieved January 5, 2010 [Available online at https://www.ncdc..noaa.gov/sotc/]

Fosberg, M.A. 1978: Weather in wildland fire management: the fire weather index.

Proceedings of the Conference on Sierra Nevada Meteorology, South Lake Tahoe,

     NV.  pp. 1-4.

Glickman, T., 2000: Glossary of Meteorology. 2nd ed. Amer. Meteor. Soc., 855 pp.

Grell, G.A., 1993: Prognostic evaluation of assumptions used by cumulus parameterizations. Mon. Wea. Rev., 121, 764-787.

Grimit, E. P., and C. F. Mass, 2002: Initial results of a mesoscale short-range ensemble forecasting system over the Pacific Northwest. Wea. Forecasting, 17, 192–205.

Haines, Donald A., 1988: A lower atmosphere severity index for wildlife fires. National Weather Digest.  13(2): 23-27.

Hamill, and S. J. Colucci, 1997: Verification of Eta–RSM shortrange ensemble forecasts. Mon. Wea. Rev., 125, 1312–1327.

Hamilton and Ostapow, 2009:  Long Island, NY - Central Pine Barrens.  National Database of State and Local Wildfire Hazard Mitigation Programs, USDA Forest Service.  [Available online at https://www.wildfireprograms.com/search.html?search=advanced&string=SUNRISE%20FIRE&state=&displayId=283].

Hamm, A., Elmore, K., 2003:  A Validation of the National Centers for Environmental Prediction’s Short Range Ensemble Forecast.  20th Conference on Weather Analysis and Forecasting/16th Conference on Numerical Weather Prediction.  [Available online at https://www.caps.ou.edu/reu/reu03/papers/Andy.pdf.

Hardy, C., R.D. Ottmar, J. Peterson, and J. Core:  2002. Smoke management guide for prescribed and wildland fire -- 2000 edition. Boise, ID: National Wildfire Coordinating Group.

Heffter J.L., 1980: Transport Layer Depth Calculations. Proceedings 2nd Joint Conference on Applications of Air Pollution Modeling. AMS, pp. 787-791.

Hoadley, J. L., K. Westrick, S. A. Ferguson, S. L. Goodrick, L. Bradshaw, and P. Werth, 2004: The effect of model resolution in predicting meteorological parameters used in fire danger rating. J. Appl. Meteor., 43, 1333–1347.

Holzworth C.G., 1967: Mixing depths, wind speeds and air pollution potential for selected locations in the United States. J. Appl. Meteorol. 6, 1039-1044.

Hydrometeorological Prediction Center, 2009:  HPC’s surface analysis archive.  [Available online at https://www.hpc.ncep.noaa.gov/html/sfc_archive.shtml].

Jones, M. S., B. A. Colle, and J. S. Tongue, 2007:  Evaluation of a mesoscale short-range ensemble forecast system over the Northeast United States.  Weather and Forecasting, 22, 36 – 55.

Kaplan, M.L., C. Huang, Y.L. Lin, and J. J. Charney, 2008:  The development of extremely dry surface air due to vertical exchanges under the exit region of a jet streak.  Meteorology and Atmospheric Physics, 102:63–85.

Keetch, J. J. and G.M. Byram, 1968:  A drought index for forest fire control.  U.S.D.A. Forest Service Research Paper SE-38.

Kassomenos, Pavlos, 2010:  Synoptic circulation control on wild fire occurrence, Physics and Chemistry of the Earth, Parts A/B/C, Volume 35, Issues 9-12, Classifications of Atmospheric Circulation Patterns - Theory and Applications, Pages 544-552.

Li, M., Qu, J., Hao, Xi, 2010:  Monitoring temporal and spatial variations of vegetation phenology from space.  [Available online at https://spie.org/documents/Newsroom/Imported/003242/003242_10.pdf]

Mesinger, F., G. DiMego, E. Kalnay, P. Shafran, W. Ebisuzaki, D. Jovic, J. Woollen, K. Mitchell, E. Rogers, M. Ek, Y. Fan, R. Grumbine, W. Higgins, H. Li, Y. Lin, G. Manikin, D. Parrish, and W. Shi, et al., 2006:  A long-term, consistent, high-resolution climate dataset for the North American domain, as a major improvement upon the earlier global reanalysis datasets in both resolution and accuracy. Bull. Amer. Meteor. Soc., 87 (3), 343-360.

Menut, L., C. Flamant, J. Pelon, and P. Flamant, 1999:  Urban boundary-layer height determination from lidar measurements over the Paris area.  Appl. Opt., 38 (6), 945-954.

Mills, G.A., 2005: On the sub-synoptic scale meteorology of two extreme fire weather days during the Eastern Australian fires of January 2003. Aust. Meteor. Mag. 54, 265-290.

Minnesota Department of Natural Resources, Section of Wildlife, 2008:  The benefits of prescribed burning on private land.  [Available online at https://files.dnr.state.mn.us/assistance/backyard/privatelandsprogram/benefits_prescribed_burning.pdf].

Mölders, N., 2008:  Suitability of the Weather Research and Forecasting (WRF) model to predict the June 2005 fire weather for Interior Alaska.  Wea. Foreast. 23(5), 953-973.

Morrin, Tim, 2009.  Interview by J. Pollina [live interview].  Observation Program Leader with the National Weather Service, New York, NY.

National Drought Mitigation Center, 2011:  U.S. Drought Monitor, published online July 6, 2011 from https://droughtmonitor.unl.edu/archive.htlm, retrieved July 7, 2011.

National Interagency Fire Center, 2008:  Fire information – wildland fire statistics.  [Available online at https://www.nifc.gov/fire_info/ytd_state_2008.htm].

National Wildfire Coordination Group, 1994:  Intermediate wildland fire behavior student workbook.  Available through the National Interagency Fire Center.

National Weather Service, WFO-New York, 174 Brookhaven Ave. Bldg. NWS-1, Upton, NY 11973.  2008:  Ready Reference Binder #7

National Weather Service, WFO-New York:  Climatological Data, published online April 7, 2008, retrieved on December 12 2010 [Available online at https://www.erh.noaa.gov/okx/climateindex.php?wfo=okx].

National Weather Service, WFO-New York:  Climatological Data, published online August 14, 2010, retrieved on May 6 2011[Available online at https://www.erh.noaa.gov/okx/climate_cms.html#Historical

Newark M.J. (1975)  The relationship between forest fire occurrence and 500 hPa longwave ridging.  Atmosphere , 13(1), 26-33.

Nimchuk N (1983)  Wildfire behavior associated with upper ridge breakdown.  Alberta Energy and Natural Resources Forest Service Report No. T/50. (Edmonton, Alberta)

NOAA National Climatic Data Center, State of the Climate: National Overview for March 2000,             published online April 2000,       https://www.ncdc.noaa.gov/sotc/?report=national&year=1999&month=15.

NOAA National Climatic Data Center, State of the Climate: National Overview for April 2000,   published online May 2000, retrieved on October 31, 2010 from         https://www.ncdc.noaa.gov/sotc/?report=national&year=1999&month=16.

Novak, D.R. and Colle, B.A., 2006:  Observations of multiple sea breeze boundaries during an     unseasonably warm day in metropolitan new york city.  Bulletin of the American Meteorological             Society, 87(2), 169-174

Parr, Dr. K., Dr. L. Instone, Dr. P. Wurm, H. Rysavy, Dr. D. Williams, L. Coleman, Dr. L. Prior, R. Mercer, R. Jalil, Dr. S. Setterfield, Dr. J Luly, and Dr. M Page, 2005:  Fire ecology and management in northern australia.  Tropical Savannas CRS and Bushfire CRC.  [Available online at https://learnline.cdu.edu.au/units/sbi263/fundamentals/fundamentals.html].

Plymouth State Weather Center, 2009:  Welcome to the Plymouth state weather center!  [Available online at https://vortex.plymouth.edu/].

Radtke, K.W.-H,  Arndt, A.M., and Wakimoto, R.H.  1982:  Fire history of the santa monica mountains.  USDA Forest Service, Pacific Southwest Forest and Range Experiment Station, PSW-58.  Technical Report.  438-443.  (Berkeley, CA)

Raphael, M.N., 2003: The Santa Ana winds of California. Earth Interactions, 7, 1-13.

Richardson, D. S., 2001: Ensembles using multiple models and analyses. Quart. J. Roy. Meteor. Soc., 127, 1847–1864.

Roebber, P. J., D. M. Schultz, B. A., and D. J. Stensrud, 2004: Toward improved prediction: High-resolution and ensemble modeling systems in operations. Wea. Forecasting, 19, 936–949.

Rorig, M.L, and Ferguson, S.A., 2002:  The 2000 fire season:  lightning-caused fires.  Journal of Applied Meteorology.  41(7), 786-791.

Saltenberger, John, 2003:  Notes on the haines index.  Fire Weather Office, Northwest Interagency Coordination Center.  [Available online at https://www.wrh.noaa.gov/sew/fire/olm/haines.htm].

Sasano, Y.  1985:  Observational study on the atmospheric mixed layer and transition layer structures using Mie lidar.  Journal of the Meteorological Society of Japan, 63, 419-435.

Schaefer VJ (1957)  The relationship of jet streams to forest wildfires.  Journal of Forestry 55(6), 419-425.

Schroeder MJ, Glovinsky M, Hendricks VF [and others] (1964)  Synoptic weather types associated with critical fire weather.  Defense Documentation Center For Scientific and Technical Information, Cameron Station Alexandria, VA.  USDA Forest Service, Pacific Southwest Forest and Range Experiment Station, (Berkeley, California).

Shein, Karsten, 2009:  Climate of 2008 wildfire season summary.  [Available online at:  https://www.ncdc.noaa.gov/oa/climate/research/2008/fire08.html].

 Skamarock, W.C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, J. G. Powers, 2005: A Description of the Advanced Research WRF Version 2. NCAR Tech Note, NCAR/TN-468+STR, 88 pp. [Available from UCAR Communications, P. O. Box 3000, Boulder, CO 80307].

Skinner WR, Flannigan MD, Stocks BJ, Martell DL, Wotton BM, Todd JB, Mason JA, Logan KA, and Bosch EM (2002)  A 500 mb synoptic wildland fire climatology from large Canadian forest fires, 1959-1996. Theoretical and Applied Climatology 71(3-4), 157-169.

Skinner WR, Stocks BJ, Martelli DL, Bonsal B, and Shabbar A (1999)  The association between circulation anomalies in the mid-troposphere and area burned by wildland fire in Canada.  Theoretical and Applied Climatology 63(1-2), 89-105.

Smirnova, T. G., J. M. Brown, and S. G. Benjamin, 1997: Performance of different soil model configurations in simulating ground surface temperature and surface fluxes. Mon. Wea. Rev., 125, 1870–1884.

Stensrud, D. J., H. E. Brooks, J. Du, M. S. Tracton, and E. Rogers, 1999: Using ensembles for short-range forecasting. Mon. Wea. Rev., 127, 433–446.

___________, D. J., and N. Yussouf, 2003: Short-range ensemble predictions of 2-m temperature and dewpoint temperature over New England. Mon. Wea. Rev., 131, 2510-2524.

Storm Prediction Center, 2009:  Surface and upper air maps.  [Available online at http://www.spc.noaa.gov/obswx/maps/].

Strawbridge, K. B. and Snyder, B. J., 2004: Planetary boundary layer height determination during Pacific 2001 using the advantage of a scanning lidar instrument, Atmos. Environ., 38, 5861–5871, 2004.

Takle, E.S., D.J. Bramer, W.E. Hellman, and M. R. Thompson, 1994:  A synoptic climatology for forest fires in the NE US and future implications from GCM simulations.  Int. J. Wildland Fire.   4(4), 217-224.

Trenberth, K.E., 1978:  On the interpretation of the diagnostic quasi-geostrophic omega equation.  Mon. Wea. Rev., 106, 131-137.

United States Forest Service, 2009:  Urbanization of forestland in the northeastern U.S. [Available online at https://www.fs.fed.us/ne/fia/studies/LDS/urban_maps/northeastmap.jpg].

University of Washington, 2009:  Electronic maproom archive.  [Available online at https://www.atmos.washington.edu/data/vmaproom/varchive.cgi].

Vogelezang D.H.P. and  A.A.M. Holtslag, 1996: Evaluation and model impacts of alternative boundary-layer height formulations. Boundary-Layer Meteorol. 81, 245-269.

Wagner, C.E. Van, 1979:  A laboratory study of weather effects on the drying rate of jack pine litter.  Canadian Journal of Forest Research9 (2), 267-275.

Werth, John, 2003:  National fire danger rating system.  National Oceanic and Atmospheric Administration/National Weather Service.  [Available online at https://www.wrh.noaa.gov/sew/fire/olm/nfdrs.htm#nfdrs6].

Wildland Fire Assessment System, 2009:  [Available online at https://wfas.net/].

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