The Four Seasons Newsletter

Welcome to the Spring 2024 newsletter as we wrap up the season. The biggest story of the spring for most was the Total Solar Eclipse! If you missed it, our Science and Operations Officer Pete Banacos wrote an  article  describing the meteorological conditions across our region during the event. We also enjoyed that night in May where a geomagnetic storm was strong enough to extend viewing of the northern lights throughout much of the US.

In this newsletter, we look back at our big April snowstorm, highlight the NWS student volunteer program, and new website that has replaced the Advanced Hydrologic Prediction Service (AHPS) for river forecast information. We also describe our partnership with the Green Mountain National Forest with an account of a prescribed burn, and take a look at the summer climate outlook. Thanks for reading and enjoy!

The last in a string of late season winter storms occurred April 3rd-5th, 2024 across Vermont and northern New York. The event produced widespread snowfall of 7 to 15 inches across the region with a few localized values in excess of 20 inches observed across higher elevations (Figure 1), and it was the most impactful of the three that took place due to its long duration and wet snow. Given its proximity to the Great North American Solar Eclipse on April 8th, this snowstorm was closely monitored for potential impacts on incoming travelers.

Forecast confidence for the event was initially high. However, confidence decreased somewhat approaching the event due to a trend towards fast mid-level flow entraining dry air that also introduced a warm layer in forecasts indicative of sleet. Additionally, near-surface conditions were expected to be above freezing for much of the event, particularly west of the Green Mountains, which were expected to limit accumulations by Thursday afternoon. Using statistics like percentiles is one way to establish credibility in a forecast, and the values remained moderately consistent leading up to the event. The 10th percentile (reasonable lowest forecast probability) indicated much of eastern Vermont and northeastern New York had high probabilities of reaching our warning criteria of 7 or more inches of snow (Figure 2). Our Winter headlines closely mirrored the values indicated in the 10th to 25th percentile. The forecast for northern New York was biased marginally high but reasonably good. Forecast snowfall amounts for Vermont were generally 4 to 12 inches too low. (Figure 3).

This event happened with different elements coming together at the right time. Cold air was present due to a strong ridge of high pressure across far northern Quebec Province and a retrograding upper low shifting from the Canadian Maritimes towards Quebec City that maintained a connection to a cool modified polar air mass. The positioning also resulted in pure divergence aloft (Top row of Figure 4). At the surface, this air was not quite cold enough to keep the boundary layer below freezing during the event. A moisture plume from the equatorial Pacific and traversing the Caribbean Sea brought anomalously high precipitable waters, but nothing extraordinary at 0.50-0.67” across the region. To our west, a large upper low near the Great Lakes developed and a deep cyclone matured over the region sending an occluded boundary north that dissipated. Along the triple point of the occluded front, a new surface low developed along the coast. This feature became the primary driver for heavy snow as it tracked near Long Island and just off the New England coastline (Bottom right of Figure 4).

Due to the strength of the Canadian high pressure and this developing coastal low, fast east-southeast winds channeled through the region. At Rutland, this was concerning due to the potential for strong downslope winds, and some forecast guidance indicated that this could have also brought in above freezing air in the mid-levels. Figure 5 shows a loop of the 850hPa mean temperature forecast and winds from the High Resolution Ensemble Forecast System (HREF).  The HREF forecasted 850hPa mean values of 70 to 80 knots across the western slopes of the southern Green Mountains of Vermont and the White Mountains of New Hampshire between about 03z and 12z on 4 April 2024, along with a small area of above freezing temperatures. Although this initially limited precipitation, strong dynamical forcing overcame these initial hurdles. Late Wednesday, April 3rd, moderate to heavy showers developed that stabilized low-level conditions and eroded the warm layer. By 12z through 18z on 4 April 2024, Vermont was located at the nose of the fast easterly low-level jet, resulting in strong speed convergence in the region.

Surface temperatures were initially well above freezing, but the combination of moderate to heavy snow and cold advection into the overnight hours helped bring several locations closer to freezing by sunrise. Then with the wind gust forecast not materializing, this likely shortened the duration of sleet for this event such that by Thursday morning on April 4th, heavy, wet snow again dominated. Figure 6 depicts the dominant precipitation types forecast in the 12z 3 April 2024 HREF run. A large area of mix was forecast through about 06z April 2024. Occasionally, pockets of light or no precipitation were observed in the HREF forecast which indicated where ensemble members depicted terrain shadowing from the strong easterly flow. Although the HREF members forecast the highest snowfall rates to lift north between 12z and 14z on 4 April 2024, heavy snow persisted through about 18z. Although conditions were expected to warm above freezing across most of the region, mid-level conditions remained cool enough for the precipitation to remain entirely as snow. However, with the lack of heavy snowfall rates, how much could accumulate was in question.

Forecast point soundings for KMPV (Figure 7) and KRUT (Figure 8) highlight the evolution of thermal profiles. At KMPV, ample cold air was present throughout the entire column. At times, the RAP forecast above freezing temperatures near the surface, but lacked a large isothermal favorable for large dendrites. The forecast soundings highlight the dendritic growth layer (DGZ) as yellow when saturated and pink when unsaturated. Forecast soundings suggested that dry air would periodically entrain in the DGZ. Additionally, the DGZ was located near or above 600hPa at the onset, with 50 to 60 knot winds below the layer. Winds of this magnitude pose the potential to shred falling dendrites. 

At Rutland, the RAP forecast suggested a warm layer reaching as high as 6 ° due to the strong easterly flow. Additionally, 70 to 80 knot winds present beneath that inversion at 850hPa were expected to produce significant downslope winds. However, this element appears to have been overdone in the RAP forecast (and others too), with a peak wind gust of only 51 mph observed at KRUT. Compared to the HREF 850hPa temperatures in Figure 5, the forecast warm layer in the RAP was notably warmer compared to the high resolution ensemble mean.

High resolution models are usually relied upon to provide more accurate depictions of terrain effects and microphysical processes. Although the HREF more accurately depicted the strength of the warm nose, it still placed too much sleet across the forecast area and its members indicated more terrain shadowing than took place. In most cases, the variations between ensemble members can account for biases. However, when every ensemble member forecasts no precipitation due to downsloping, it negatively impacts the forecast of events like this.

Synoptic weather models were effective for this event. However, the 3 to 6 temporal resolution they are often in can make it challenging to analyze the behavior of meso bands and subtleties in low pressure tracks. In this instance, the surface low forecast during the most important part of the coastal low’s development from the Weather Prediction Center was excellent for this event, and the synoptic guidance proved consistent (not shown). Strong model consensus was present even 5 days before the event (Figure 9). Ensemble clusters based on their forecast conditions at 500hPa and the whole collection known as the “Grandensemble” indicate moderate to heavy snow in most clusters. Only cluster 4 had a 500hPa configuration creating an appreciable difference in the snowfall forecast. In this situation, using synoptic weather models to minimize the effect of terrain shadowing in high resolution guidance would have been reasonable given the consistent snowfall forecasts.

Bringing each piece together, dry air ahead of the coastal low, the expectation of a sleet mix and terrain shadowing, and how quickly the mesoband was depicted to lift north, each expected forecast from the NWS Burlington office was generally near the 10th and 25th percentile. Compared to the box and whiskers depicted in Figure 10, the observed snowfall totals roughly fell between the median and the 75th percentile values across Vermont. Observed snowfall amounts for northern New York ranged between the 10th and 25th percentile, though. Although the percentile value that was most characteristic for each city varied, none were outside the range provided by model guidance.

At the beginning of the event, precipitation struggled as the occluded front dissipated and pockets of dry air lifted north. However, the dry slot never completely wrapped into the circulation of the original upper low, and rather translated towards the east as an embedded circulation began to develop. The water vapor imagery in Figure 11 shows the development of the curl moving north from the Mid-Atlantic northeast overhead. Another curl took shape as it lifted north towards Vermont at about 15z, and would continue to produce heavy snowfall across the region into the afternoon hours on 4 April. Dry air remained on the eastern side of this developing upper level feature that stalled over the region at the end of the loop. This prevented any additional dry air from reaching the area. Additionally, the noted speed convergence and isentropic lift within a TROWAL (trough of warm air aloft) resulted in another burst of snow across Vermont at the end of the loop.

Viewing the evolution from radar, intervals of precipitation were interrupted by the dry mid-level conditions. Precipitation shifted into the region Wednesday afternoon on 3 April 2024 as an occluded boundary shifted into Vermont and northern New York, but stalled and dissipated. However, late in the evening, a subtle zone of 700-500hPa frontogenesis developed ahead of the developing coastal low (not shown), which produced moderate to heavy precipitation. Even in some of the lower valleys, snow began to mix in with this wave that lifted north between about 8 PM and midnight. This occurred when the strongest wind gusts were expected to develop and may have contributed to the weaker downslope winds that followed. There was a brief wintry mix with mainly sleet here at BTV during the overnight hours (mainly from 11 PM till around 1 AM), which led to a slushy coating on elevated surfaces. Then there was a 3 hour lull in precipitation and the temperature rose to 37 °F. Then by 4 AM on April 4th, snow overspread the region as the coastal low developed and the main event began (Figure 12).

Snow remained prevalent through the morning hours across the region. The general expectation was for the swath of moderate to heavy snow to quickly lift north of the international border and for snowfall rates to decline mid-morning (Figure 6). A decrease in snowfall rates would have also allowed temperatures to warm above freezing given the time of year. However, the region of moderate to heavy snow was slower to lift north than anticipated. Rather than make a quick exit, snow even began to pivot back into the region late in the morning, with the 0 to 12 km hodograph showing the favorable northeast to southwest winds common in such events. With the low level jet shifting east, speed convergence across the area combined and steep mid-level lapse rates of 7 to 7.5 °C/km across eastern Vermont maintained moderate to heavy snowfall rates longer than the HREF had anticipated (Figure 13). By mid-afternoon, snowfall rates declined but precipitation types never transitioned back despite warming above freezing. Snow continued into the overnight hours Thursday into Friday.

The downslope wind conditions offered a complicating factor to the forecast. In most situations, easterly flow will advect warm air into the region, and subsidence results in drying that precludes blizzard conditions. To meet blizzard criteria, three conditions must be met. There must be heavy or blowing snow that drives visibility to a quarter mile or less, there must be frequent wind gusts in excess of 35 mph, and these conditions must last for 3 consecutive hours. Wind gusts of 60 to 70 mph, locally up to 85 reported by a spotter in Petersburgh of Rennselaer County, were observed to our south. However, for our region, observed gusts were mainly below 50 mph outside of a few favorable downslope locations (Figure 14).

The western slopes of the Greens were the most likely region to experience localized blizzard conditions. The potential for blizzard conditions in mountain passes was conveyed in forecast messaging. Rutland lies along the foothills of these slopes and is prone to east or southeast downslope winds. Figure 15 shows wind speeds, gusts, and visibilities with the threshold for blizzard winds and visibilities marked. Between about 00z and 03z (8 PM and 11 PM EDT), Rutland had frequent gusts but lacked low visibility, and then after 06z (2 AM EDT) was close on visibility but not the gusts. Given the strength of the 850hPa flow at 09z in Figure 16, the snow helped stabilize gusts at lower elevations. The intermittent precipitation makes it difficult to verify whether blizzard conditions occurred, but travel conditions were certainly dangerous through mountain gaps.

A powerful storm capped off the cold season with one final event with heavy snow between 3-5 April 2024. Despite forecast guidance struggling to depict the pivoting mesoband, their forecast accumulations were good (consistent and within probabilistic ranges) for this event. Intrusions of dry air initially caused concern that the event might not reach the 25th percentile of forecast accumulations. However, a surge of heavy precipitation late 3 April 2024 helped to stabilize conditions and bring an early end to sleet, followed by heavy snow lingering later than forecast, that combined to produce another late season snowfall event ranked high for this time of year. Several COOP sites observed 24-hour accumulations in the top 5, and our Corinth, VT observer measured their highest 24-hour April snowfall accumulation (Table 1). This snow was wetter and resulted in a prolonged period of outages that peaked early 4 April 2024 with nearly 35000 without power across Vermont alone. Fortunately, dry air on the western periphery of the departing upper low produced pleasant weather conditions in time for the Great North American Eclipse days later.

Overall, this was a well forecasted and messaged winter storm, including the heavy snowfall rates coinciding with the morning commute. Despite temperatures staying above freezing across the Champlain Valley for the entire event, the sheer synoptic and mesoscale dynamics allowed the heavy snowfall rates to persist past the early morning commute across the Champlain Valley. The operations team used various tools to get out the message, including an hourly snowfall loop on the email brief, Slack and social media platforms. The team was also proactive in issuing a SPS for heavy snowfall rates during the morning commute, as well as upgrading to a Winter Storm Warning as the heavy snowfall rates showed no signs of letting up.

The National Weather Service Forecast Office in Burlington, VT (BTV) provides volunteer opportunities for students considering careers in meteorology and hydrology. During the spring 2024 semester, NWS Burlington partnered with the University of Vermont (UVM) College of Arts and Sciences, and the Office of the Vermont State Climatologist, to host UVM junior Miya Whitfield. 

Throughout the semester, Miya had the opportunity to learn about NWS operations and assist BTV forecasters with operational duties. In addition, Miya completed a local research project entitled, “Placing Vermont’s July 2023 Precipitation into a Climatological Context”. Reflecting on her experience, Miya conveyed, “It was a great experience to both observe the operations side of the office and contribute to research on a locally relevant topic. Having had this opportunity, I now know that I am interested in pursuing further studies in atmospheric science.” NWS Burlington looks forward to continued collaboration with UVM and other area colleges and universities to enhance student learning and career-related experiences.  

For comprehensive and up-to-date information on student opportunities throughout the NWS, please see the Student Opportunity StoryMap below:

On 26 April 2024 two meteorologists from the National Weather Service Office in Burlington, Vermont were deployed to provide weather support to the Green Mountain National Forest for the Perry Basin Prescribed Burn. A prescribed burn is controlled fire application by a team of fire experts performed under specified weather conditions to help restore health to an ecosystem that depends on fire. This prescribed burn took place on the Green Mountain National Forest near Granville, Vermont at a midslope elevation of 2000 feet and was about 26 acres.  

We arrived at the Rochester Ranger Station to meet the task group of firefighters with the Forest Service assigned to the burn, while learning the burn objectives and any potential hazards. In addition, we discussed the weather support needed, which included an incident weather briefing on expected weather conditions, along with providing half hour weather updates on the radio. The half hour weather updates included current temperature, relative humidity, wind direction, wind speed, fine dead fuel moisture, and the probability of ignition based on the conditions, along with any significant wind shifts, changes in fire behavior, and smoke observations. 

The synoptic scale weather featured large and strong high pressure directly overhead, which provided the prescribed burn site with plenty of sunshine and mostly light upslope terrain driven winds. Winds were mainly from the east-southeast at 2 to 4 mph with a few gusts 4 to 6 mph that occurred around 1400. As drier air mixed toward the surface we saw a minimum relative humidity value of 20% at 1500. Humidity values ranged between 20-24% during the burn and started to improve toward 1700. 

For more than two decades, the National Weather Service (NWS) delivered water resource forecast products and services though the Advanced Hydrologic Prediction Service (AHPS). This legacy system needed to be modernized to support newer innovation with technology supported by scalable infrastructure that uses services-driven architecture.

The NWS has developed a new gateway to water resources forecasts and information on the web, called the National Water Prediction Service, or NWPS ( water.noaa.gov ) which was just released on May 27, 2024. The new NWPS interface improves on the old AHPS system through better web functionality and a modern, more efficient mobile-friendly web code. NWPS is now the primary source of NWS water data and hydrological forecast information and includes improved access to real-time river information, precipitation estimates and more overall meteorological and hydrological data. The website shown above links to the national scale after which you can zoom into your area of interest. However, customizable plots by state are also readily available. For example, for a zoom only for Vermont go to  water.noaa.gov/state/vt , and for New York,  water.noaa.gov/state/ny .

Viewers will have the most up to date information, in a fraction of the time. The gauge hydrograph views utilize dynamic, customizable plots generated on the user’s web browser to enable more speed and timely data for users. Additionally, users can now visualize data as far back as 30 days rather than 7 days under the legacy AHPS interface. This allows for analyzing trends and patterns that are crucial for better understanding water resource management, flood forecasting and environmental monitoring.

NWPS gives us the first look at the National Water Model (NWM) on the scale of individual river gauges and stream reaches. This model data provides initial guidance values for all rivers and streams in the nation. As one zooms in, more of the network is shown.

The FIM capability developed under the legacy AHPS system has transitioned to NWPS as a custom local web map from the gage hydrograph view page. It provides the same visualization as the legacy AHPS static FIM libraries. At this time, FIM services are available for 10% of the United States population, which includes portions of New York and far southern Vermont. By October 2024, FIM will be available for the remainder of Vermont and all of Northern New York. For more on the FIM roll-out plan, please visit this  link .

In addition, all NWPS data is available via the new NWPS API. This allows users to include NWPS information directly into their own applications and services.

**For a full overview of the NWPS interface’s capabilities and functionality, check out this StoryMap:

Footnotes:

 Accessibility Statement 

  Privacy Statement