Halifax 2010 Lidar Measurements Archive

This is preliminary data. For more information, contact:

Prof. Thomas J. Duck
Department of Physics and Atmospheric Science,
Dalhousie University,
Halifax, Nova Scotia, Canada, B3H 3J5

Tel: (902)494-1456
Fax: (902)494-5191
E-mail: tom.duck@dal.ca

Halifax is located at 44.6N, 63.6W.

Please see the examples to learn how to interpret our measurements.

The following additional data sets are available:



2010-07-03 03:35:21 to 2010-07-03 23:30:22

Light aerosol loading is found below 2 km altitude, with an interesting aerosol cloud observed between 0800 and 2000 UTC between 2 and 3 km. Cirrus clouds can be seen after 1200 UTC above 10 km altitude.


Diagnostic plots: +
More plots: +

2010-07-04 14:05:00 to 2010-07-04 22:29:40

This day shows stratified aerosol layers below 2 km altitude, with a surface fog bank near the end of the measurement. Thin cirrus at 12 km altitude are evident during the first half of the measurements. After 1900 UTC the cirrus thicken to the point that only the lower boundary is detected (shown in white), and no signal is obtained above it (shown in black).
There seems to be slight contribution from biomass burning sources to the aerosol layer detected. The fires across North America during the previous week do inidcate that they may be the point of supply of this boreal biomass CO to the plume at an altitude of 2km.


Diagnostic plots: +
More plots: +

2010-07-05 20:52:06 to 2010-07-06 05:59:39

Stratified aerosol layers appear below 2 km altitude, following similar observations the previous day. The thin red layer below 1 km altitude is likely thin cloud/fog. Notice that between 0000 and 0100 UTC on July 6 it became optically dense enough to extinguish the laser beam -- hence the enhanced noise at higher altitudes during that time. The bottom of a cloud deck with varying altitude is seen above 6 km altitude.


Diagnostic plots: +

2010-07-07 11:14:18 to 2010-07-09 22:59:58

This measurement, spanning three days, reveals aerosol plumes below 1 km altitude. Fog rolled in around 12 UTC each day, and data gaps -- in black and leading to noise at high altitudes -- resulted. Cirrus were not observed until the final hours of the measurement.


Diagnostic plots: +

2010-07-13 18:37:06 to 2010-07-13 21:35:02

Conditions only permitted a short measurement on this day. The bottom of a low cloud deck is seen between 4 and 6 km altitude. Stratified boundary-layer aerosols were found below 3 km altitude.


Diagnostic plots: +
More plots: +

2010-07-15 11:10:55 to 2010-07-16 13:15:46

This measurement spans 24 hours over two days. Light aerosol loading is seen below 2 km altitude. Most interestingly, local boundary-layer development can be seen between 1400 and 1900 UTC on July 15. Clouds appear at the top of the boundary layer thereafter, often causing retrieval artifacts at lower altitudes (these will be corrected in due time). Patchy cirrus clouds are apparent above 8 km altitude.

FLEXPART Comparison

FLEXPART modelled dispersions from forest fires in Central Canada show a plume of emissions in a similar location to the one measured by the LIDAR. The chemical species used as a tracer of burning emissions for this model was Carbon Monoxide. Using instead an artificial species CO-wdep, which acts more like a hydrophillic aerosol, shows that some of the plume has been removed, though the bulk of it remains. This suggests that the aerosols in the plume measured by the LIDAR are mostly sourced from fires in Central Canada. These emissions took 4 to 5 days to reach Halifax.


Diagnostic plots: +
More plots: +

2010-07-17 16:32:02 to 2010-07-17 23:44:16

This measurement only spans over the evening of July 17. Below approx. 2km altitude dense aerosol coverage is observed. The thick plume at around 1700UTC is most likely pollution from major cities along the east coast of the USA, as indicated by both the HYSPLIT back trajectory and the MODIS hot spot image. The denser part of the plume between 2200 and 2300UTC is likely from forest fires in northern Saskatchewan (View HYSPLIT).


Diagnostic plots: +
More plots: +

2010-07-20 14:12:01 to 2010-07-20 16:52:45

The measurement from July 20th, only last a few hours. Throughout the first hour there is a visible aerosol layer with a maximum altitude of ~2km around 15:30. A HYPSLIT back trajactory has been run for this time at three differant altitudes, as well as a HYSPLIT Matrix for the adjacent area around Halifax. To examine the potential sources of the aerosols, its possible to relate to the MODIS Hotspot or the Environmental Canada fire website.
GEOS-5 forecasts clearly indicate a layer of CO aerosol, but there is no indication of it being from biomass sources.
There is visible cloud coverage with a large optical depth evident throughout the majority of the measurement at an altitude of 8-9km.


Diagnostic plots: +
More plots: +

2010-07-21 06:54:29 to 2010-07-22 07:12:13

This measurement spans 24 hours over two days. A nice example of the approach of a cold front can be seen starting with some high level cirris clouds at about 1600 UTC on the 21st. This layer of clouds gradually thickened before starting to block out the entire laser beam and begining to descend at around 0300. The frontal rain belt finally arrived at 0715 on the 22nd, and the measurement had to be stopped.

Also clearly visible is a layer of aerosol which arrived above the Lidar at around 1200 UTC on the 21st at an altitude of around 3km. As the aerosol plume was transported overhead it descended to around 1km before dissapearing at around 0400 on the 22nd. The plume's altitude, shape, and arrival time agree very well with a geos-5 forecast of high Carbon Monoxide levels due to Boreal Biomass Burning that was produced by Mark Parrington at the University of Edinburgh.

Running a hysplit back trajectory for 120 hours (5 days) ending at 1300 UTC at 2800m height on the 21st shows that the air at that point may have come from British Columbia. And indeed looking at the fire hotspots map produced by Natural Resouces Canada for that day (the 16th) shows that there were significant forest fires in that region, suggesting that this may have been the source of the aerosols. However, the hysplit trajectories do not provide any kind of quantified measure of the uncertainty. To remedy this a set of 9 Hysplit backtrajectories, with end points arranged in a grid centered on Halifax, was also produced. The limit for this sort of calculation is a 72 hour plot, but it is clear that on the 21st Halifax appears to have been at the focus of converging wind patterns, suggesting that there is a large uncertainty in any single trajectory. For this reason it is clear that the smoke plumes could realistically have come from anywhere in Northern or Western Canada, for example the fires from the previous few days along the province boundaries of Nunavut, Manitoba and Saskatchewan.

FLEXPART Comparison

This interpretation seems appropriate given the FLEXPART modelled dispersion from forest fires in Central Canada, which shows a plume of emissions in a similar location to the one measured by the LIDAR. The chemical species used as a tracer of burning emissions for this model was Carbon Monoxide. Using instead an artificial species CO-wdep, which acts more like a hydrophillic aerosol, shows that some of the plume has been removed, though the bulk of it remains. This suggests that the aerosols in the plume measured by the LIDAR are mostly sourced from fires in Central Canada, and not from British Columbia. These emissions took 7 to 9 days to reach Halifax.


Diagnostic plots: +
More plots: +

2010-07-23 05:31:27 to 2010-07-23 23:57:00

This measurement runs for a duration of just over 18.5 hours. From the Lidar plot, during the first ~6 hours of the measurement there seems to be 3 distinct aerosol layers present at 2km, 4km and 6km. This is verified with the back trajactories that show three paths, with similiar routes for the two higher altitude trajactories.
The GEOS-5 forecast output indicates that the abundance of the aerosol layers present at 2km is from a biomass burning source. The forecast indicates that the aerosol measured around 6km is also from a biomass burning source, however this is not as prevalent. The MODIS image shows the fire hotspots for 5 days before the 23rd. The aerosol from suspected biomass burning sources for the 23rd tends to end around 15UTC.
The last 7 hours has been affected by low level cloud.

FLEXPART Comparison

FLEXPART modelled dispersions from forest fires in Central Canada show a plume of emissions in a similar location to the low level one measured by the LIDAR. The chemical species used as a tracer of burning emissions for this model was Carbon Monoxide. Using instead an artificial species CO-wdep, which acts more like a hydrophillic aerosol, shows that some of the plume has been removed, in particular the high level aerosols seen by the LIDAR up to 8km are not modelled when deposition of the aerosol is considered, this might indicate that FLEXPART had a deposition rate that was too high, or that these aerosols have a source other than Central Canadian forest fires. The large lower plume is modelled, though it has it's maximum concentration at a time when the LIDAR image was being affected by low level cloud, so it is difficult to know whether the model agrees with the measurements or not. The Central Canadian emissions took 9 to 10 days to reach Halifax.


Diagnostic plots: +
More plots: +

2010-07-26 04:10:41 to 2010-07-28 07:59:33

This measurement spans for three and a half days from early on July 26 to noon on July 29. Due to the length the plot was split in two, this being the first part. There were a few low lying clouds around 2000UTC on July 26 and 1800UTC on July 27. There is some very optically thin aerosol present and using HYSPLIT back trajectories in conjuction with MODIS images we can determine their approximate origin.

FLEXPART Comparison

FLEXPART modelled dispersions from forest fires in Central Canada show a large mass of emissions that has not been measured by the LIDAR. The chemical species used as a tracer of burning emissions for this model was Carbon Monoxide. Using instead an artificial species CO-wdep, which acts more like a hydrophillic aerosol, shows that one explanation for the discrepency is that almost all of the aerosol that were associated with the emissions appear to have been removed, so that the LIDAR cannot measure them. The strongest of the remaining modelled aerosol plumes appears to be conected with a measured plume at 6km in the last few hours of the measurement. The sources of the other weak high altitude plumes is not clear The 6km plume took under 5 days to reach Halifax.


Diagnostic plots: +
More plots: +

2010-07-28 07:59:34 to 2010-07-29 11:13:24

This is the continuation of the measurement started on July 26. On July 28 around 1500UTC at approximately 12km altitude there is a very dense plume. HYSPLIT backtrajectories suggest it to originate form the west coast of USA. MODIS however show no significant forest fires at the appropriate time, it would take some very powerful pyro convection to lift a plume to that height anyway so it is likely that this one has another source, other explanations could be volcanic ash or simply a very thin cloud, we have yet to disprove flying spaghetti monsters as a hypothesis. At lower altitudes, but roughly the same time, we observe the lower boundary of an aerosol layer steadily decreasing in height with time. This layer exists in air that has past over Central Canada (View HYSPLIT), however FLEXPART suggests that fires in Eastern Siberia were the source of the aerosols. On July 29 at 0300UTC and about 1km altitude the aerosol becomes thicker than previously. HYSPLIT images and MODIS show that this plume is in air that has travelled from Central Canada by way of the industrialised East Coast of the United States. The dense clouds that came in at 0800UTC on July 29 extinguished the beam and the measurement was stopped.

FLEXPART Comparison

FLEXPART modelled dispersions from forest fires in Central Canada shows a large mass of emissions throughout the period in the lower 3km of the atmosphere, with another distinct plume at 5km during the morning of the 28th. However correcting for any aerosols present using the method described in previous posts leaves only the upper level plume, which is barely visible on the LIDAR image, and a large lower plume on the afternoon of the 29th, which does agree with LIDAR measurements of very thick aerosol at that time. Therefor this thick plume appears to be composed at least partly of emissions from Central Canada which took took 6 to 7 days to reach Halifax.

FLEXPART modelled dispersions from forest fires in Eastern Siberia shows a thin descending plume begining around 4km on the morning of the 28th, and descending to 3km by midnight of the 29th, afterwhich there is a large thick plume in the lower 3km until the end of the plot. This is in remarkably close agreement to the LIDAR measurement. Modelling with correction for aerosol deposition removes some of the plume but retains the basic shape, and the similarity of the shape of the original model to the measurements suggests that in this case we can be confident that FLEXPART has had too high a deposition rate, and that the descending plume that was measured is composed of emissions from Eastern Siberia. It's also clear that the large lower level plume also contains some aerosols from fires in Eastern Siberia, which took 10 to 12 days to reach Halifax.

The aerosols that were measured in the lower 2km prior to midnight on the 29th do not appear to have a forest fire source, and these, plus probably a component from the thicker plumes later, may have come from anthropogenic sources in the Eastern United States.


Diagnostic plots: +
More plots: +

2010-07-31 05:53:43 to 2010-07-31 14:54:11

The measurement on July 31 lasted for roughly 9 hours. Between 1000UTC and 1100UTC very low altitude clouds, quite possibly dense fog, extinguished the beam entirely. Low hanging clouds continued for several more hours making it hard to normalize the beam, creating messy data. There does appear to be a thin aerosol layer at 2km for the first 4 hours of the measurement, and some boundary layer aerosol below 1.5km.Beyond 1000UTC it is impossible to tell what aerosol exists.

FLEXPART Comparison

The FLEXPART modelled dispersion from forest fires in both Central Canada and Eastern Siberia show plumes situated within the lower 4km of the atmosphere during the time span of this measurement, though modelling with a hydrophilic aerosol ( Central Canada, Eastern Siberia suggest that the absence of these plumes in the LIDAR image is due to deposition of the aerosols. The fog and low cloud that disrupted the image suggest that there was a lot of moisture present in the air, and this would be expected to lead to faster deposition of the aerosols.


Diagnostic plots: +
More plots: +

2010-08-01 16:50:31 to 2010-08-03 20:25:50

For the measurements from August 1 to August 3 we observe a continuous aerosol layer for heights up to 2km, and clouds of thinner aerosol elsewhere up to heights of 9km. Despite the presence of fog on the morning of August 2 we were still able to continue collecting data until the evening of August 3 when dense clouds and rain rolled in. HYSPLIT1: August 1 2100UTC (HYSPLIT2: August 2 1200UTC) show the direction of the prevailing winds that at the levels of the aerosols.

FLEXPART Comparison

The FLEXPART modelled dispersion from forest fires in both Central Canada and Eastern Siberia show plumes situated within the lower 4km of the atmosphere during the time span of this measurement, though modelling with a hydrophilic aerosol ( Central Canada, Eastern Siberia suggest that the low quantities of these plumes in the LIDAR image is due to deposition of the aerosols. The fog and low cloud that disrupted the image suggest that there was a lot of moisture present in the air, and this would be expected to lead to faster deposition of the aerosols. The thick aerosol that is being measured in the lower 2km might contain some contribution from these fires but are likely to be dominated by other sources, possibly local polution.

The thinner higher plumes that are measured may be the same plumes that are modelled for emissions from Eastern Siberia, unfortunately high level clouds obscure much of the upper troposphere so that it is difficult to see how well the positions correlate to what was there. These plumes have been completely removed by modelling a partical that experience deposition (using the method explained in previous posts), though this is likely to do with too high depositional rates being parameterized in FLEXPART.


Diagnostic plots: +
More plots: +

2010-08-06 14:22:11 to 2010-08-08 12:00:19

Heavy surface aerosols (below 4 km) are seen from beginning of measurement to 1200 UTC. HYSPLIT trajectories show that this most likely comes from pollution in the NE Untited States.

A particularly dense aerosol cloud is seen on Aug 8 from 0000 UTC to 0008 UTC from 4-8 km. HYSPLIT trajectories show that this air comes from the west across northern Canada. MODIS hotspots show that there were fires in northern Saskatchewan in the days prior to the plume.


Diagnostic plots: +
More plots: +

2010-08-08 12:00:20 to 2010-08-10 05:04:18

This measurement spans about 40 hours and shows low level aerosols throughout. There is a cloud layer just above 2 km that disrupts aerosol observation, as well as a cloud layer at 8-10 km.

HYSPLIT models show that the air from 0-2 km comes from NE US and Northern Saskatchewan. MODIS hotspot plots show fires in northern Saskatchewan, indicating that some of the aerosols might be from boreal forest fires there.


Diagnostic plots: +
More plots: +

2010-08-10 19:46:58 to 2010-08-11 07:29:42

This measurement spans 12 hours and shows low level aerosols throughout. Clouds starting at 0200 UTC Aug 11 disrupted aerosol observation. HYSPLIT trajectories indicate that the aerosols originated in Northern Saskatchewan, where there is a high concentration of MODIS hotspots.


Diagnostic plots: +
More plots: +

2010-08-11 16:41:47 to 2010-08-13 00:30:07

This measurement runs about 32 hours. There are significant aerosol plumes at 4-8 km, with the highest concentration occuring at 0000 UTC on Aug 12. HYSPLIT shows that these aerosols come from northern Quebec, but MODIS does not show any hotspots in this region in the days prior.


Diagnostic plots: +
More plots: +

2010-08-13 00:30:08 to 2010-08-15 00:01:25

This measurement runs just over 48 hours. Cloud cover at 3 km interrupes the measurement from 1500 UTC Aug 13 to 0000 UTC Aug 14.

Aerosols are seen descending from 6 km at 0000 Aug 13 to 3-4 km at 1800 Aug 14. HYSPLIT trajectories show that all aerosols originate over Northern Quebec. MODIS shows no hotspots along the trajectory path.


Diagnostic plots: +
More plots: +

2010-08-15 00:01:26 to 2010-08-16 19:04:00

This measurement is just under 2 days long with significant aerosol layers and boundaries visible. Most of the aerosol is present for up to 3km altitude. As seen by the combinationg of HYSPLIT1 (on August 15 at 0200UTC), HYSPLIT2 (on August 15 at 1000UTC), HYSPLIT3 (on August 16 at 0500UTC) and MODIS images the most likely source of these distinct aerosol layers is Cape Breton and PEI. Measurements were discontinued due to fog immediately followed by low dense clouds.


Diagnostic plots: +
More plots: +

2010-08-20 15:00:21 to 2010-08-23 19:53:42

This measurement spans three days. Notable aerosol events include surface level aerosols, strongest from 1200 UTC Aug 20 to 1200UTC Aug 21 and 0600 UTC Aug 23 through the end of the measurement, and an aerosol plume at 6-8 km from 0600 UTC to 1800 UTC Aug 21. HYSPLIT trajectories indicate that the surface aerosols come from pollution in the northern US, whereas the high-altitude plume comes from burning boreal forests in Northern Saskatchewan which are seen as MODIS hotspots .


Diagnostic plots: +
More plots: +

2010-08-24 14:07:54 to 2010-08-26 00:48:49

This measurement spanned 30 and a half hours from August 24 to August 26. Isolated aerosols can be seen descending from 4km to 3km over the course of 24hrsfrom 1800UTC on August 24. Low clouds are scattered throughout the measurement (altitude < 2km) and interrupt the data from 1500UTC through to 1900UTC on August 25. High altitude clouds were detected at ~12km that same day. The measurement was terminated due to rain early on August 26.


Diagnostic plots: +
More plots: +

2010-08-27 13:26:44 to 2010-08-31 04:13:05

This measurement spans two and a half days from August 27 to August 31. There are minimal clouds and light aerosols <3km on August 29-30. There is a denser aeosol event starting on August 30 in the same altitude range and continuing into August 31.


Diagnostic plots: +
More plots: +

2010-08-31 04:25:13 to 2010-08-31 23:52:35

This measurement was a continuation of the previous measurement spanning 19hrs on August 31. The light aerosol layers continue <3km for the majority of the day. Denser aerosol is observed starting at 1800UTC. High altitude clouds can be seen at ~10km.


Diagnostic plots: +
More plots: +

2010-09-02 13:55:48 to 2010-09-02 21:21:45

This was a short 8 hour measurement on September 2. It was taken during the Halifax heat wave preceding Hurricane Earl (landfall September 4). A dense, low-altitude layer of aerosol can be seen <2km with clear skies above. The density and persistency of the aerosol layer was likely affected by both the heat and incoming storm system. The path of the aerosol as modelled using HYSPLIT to show a potential orgin along the Eastern US coast.


Diagnostic plots: +
More plots: +

Your request took: 1.39 seconds