Monthly Archives: February 2014

Local Meteorology and Air Quality

During certain periods, the air here at Bodega Marine Lab (BML) has been stunningly clean — approaching what might be found in the remote ocean.  We’re finding, somewhat unsurprisingly, that local meteorology heavily affects the composition of particles that we sample at BML.

Here’s a case study from last weekend.  The scene: nothing but Pacific Ocean to our west, with cow pastures, campgrounds, and a popular vacation town to our east and south.  On sunny days under fair weather (high pressure) we experience a diurnal cycle of sea breeze and land breeze.  In aerosol particle terms, we see a very segmented signal: distinct ‘sea breeze’ and ‘land breeze’ aerosol characteristics.  Let’s have a look at the aerosol particle size distributions next to the surface meteorological conditions (from the nearby DRI met station on Bodega Marine Reserve):


Two days of aerosol particle size distributions (top two panels), carbon monoxide and solar radiation (3rd panel from top), ambient temperature and RH, with wind speed and direction on the bottom panel. The correlation of wind direction with the number concentration of small particles (2nd panel from top) is striking.

The sources of particles in air masses that we sample during the sea breeze (wind from 300 degrees) are ideally very limited: particles ejected from the ocean as sea spray, and particles that form through chemical reactions in the atmosphere just above the ocean.  In contrast, the land breeze (wind from about 140 degrees)  carries with it particles from a wide variety of sources: fossil fuel combustion, biomass combustion (fires), food cooking, and agriculture.

If you aren’t familiar with this kind of data, and you have a keen eye, you’ll start to figure out what’s happening just by thinking about what particle sources we’re sampling under the land breeze vs the sea breeze.  Since the sea breeze contains sea spray, you can see that sea spray aerosol particles are typically characterized by larger particle sizes (the colors in the top plot show higher concentrations near the 1 micron diameter mark), compared to the kinds of aerosols observed during the land breeze episodes (smaller proportions of larger particles compared with to the huge number of small particles).  It’s well known to atmospheric chemists that particles that are emitted directly from combustion sources (so-called ‘primary’ combustion aerosol) are very small, and tend to grow as they react with other trace gases in the atmosphere or are incorporated into clouds and fogs.

I can tell you anecdotally, that the composition of the particles is very different between these two episodes.  Sea salt particles (containing mainly sodium chloride) that travel through polluted air quickly react with the nitrogen oxide pollutants to form sodium nitrate particles.  The land breezes also bring with them fresh combustion particles.  You’ll notice some small spikes in carbon monoxide on either end of the sea breeze periods, just before and after the switch — those are actually from fires from local campgrounds directly upwind of us — it was a weekend in a vacation destination town after all!

One really exciting thing for a atmospheric chemist is to see their science in the world around them.  To live in and around the environment; to observe with your own senses what you’re observing with your instruments.  This is why I’m an environmental scientist.

-Doug Collins, Grad Student (UCSD)

Round 2 is on its way

Greetings from Bodega Head!

After a good soaking yesterday, the sun has come out for a brief hello.


Recap of Wednesday’s rain (round 1)

We received around 2 and a quarter inches of rain here yesterday, largely between 7 am Wednesday morning and 8 pm Wednesday night. A frontal boundary passed us around 7 pm Wednesday. It wasn’t a classical cold front, which can be seen from the 449 MHz wind profiler and RASS timeseries:


Time-vertical cross-section of wind and virtual temperature from 449 MHz radar + RASS at BBY for the period ending at 11 am PST on Feb 27, 2014. Credit: NOAA-HMT.

As can be seen in the wind profiles, winds turned from SSE to SW after the boundary, and the lower atmosphere actually warmed. As we will see, freezing levels did not change much after the boundary either.

The profiling radars at Cazadero (CZC) and Santa Rosa (STR) capture this beautifully:


Time-Vertical cross section of signal-to-noise ratio from the S-Band Radar at CZC for the period ending at 11 am PST on Feb 27, 2014. Credit: NOAA-HMT.


Time-vertical cross section of signal-to-noice ratio from the S-Band Radar at STR for the period ending at 11 am PST on Feb 27, 2014. Credit: NOAA-HMT.

For each case, the bright band does not ascend or descend after the time of frontal passage. It is apparent that CZC received a period of very heavy rain between 6 pm and 8 pm, near the frontal boundary. In fact, cloud echo-top heights reached to nearly 9 km AMSL. The wind profiler here at BBY shows that this period coincided with a dramatic increase in upsolpe water vapor flux:


Time-vertical cross section of wind and timeseries of upslope IWV at BBY for the period ending at 11 am PST on Feb 27, 2014. Credit NOAA-HMT.

In this case, the upslope wind direction is defined along the vector which points from BBY to CZC. as Southerly winds really sped up near 6 pm yesterday, upslope IWV also shot up, and for the period where the value exceeded 20 knot-inches, precipitation was enhanced at CZC compared to BBY.

Potential for dust in precip. samples:

Yesterday in this post, I mentioned that we observed dust at the surface for most of the morning, while the storm was coming ashore. We also expected high altitude dust in the storm’s central circulation. There are a few ways that elevated dust could reach our site at the ground. 1) Dust aerosol could be scavenged by falling raindrops, which then make it into our precipitation sampler. 2) Dust aerosol could act as cloud condensation nuclei, which grow into or become incorporated in a rain-drop or precipitating ice and make it into our precipitation sampler. 3) Dust aerosol could act as ice nuclei, which grow into ice-phase precipitation particles, melt on their path to the surface and make it into our precipitation sampler.

We are most interested in the last possibility, which means that we would really like to collect dust during periods when the precipitation profiling radar is telling us that there is cloud ice over our heads (i.e. a brightband is present) in a deep layer.

Even at STR, whose radar likely mimicked the cloud structure that existed over us at BBY, there were long periods of high cloud-top echoes over brightband rain yesterday. Given that we expected a deep tropospheric layer of Asian dust, particularly in the early phases of this storm, that is good news for our precipitation residue analysis!

Looking ahead to Round 2:

Since late yesterday evening we have entered an interlude between storms. Winds were southwesterly for a short time, but will soon become southerly in advance of the next system. We are going to experience a rare event here on the Northern California coast. We are going to see a small but powerful extratropical cyclone whose circulation center passes to our south. We expect the low pressure center to come onshore early Saturday morning near Monterrey Bay. Before this time, we will encounter another frontal boundary here at BBY. According to model forecasts, a mesoscale band of strong SE winds will cross the coast near 4 am PST tomorrow morning:


10m wind speed and wind barbs forecast by 4 km MM5 valid at 4 am PST on Feb 28, 2014. Credit: DRI-CEFA.

After this time, winds will shift to SE, then rotate to S, then E, then eventually NE and Northerly as the low pressure center comes fully onshore. We expect high wind speeds at several points tomorrow, as the particular series of events mimics a rainband, then near-miss of a tropical storm. Pretty unique for the west coast!

The bottom line is that it should start raining again tonight at around 4 am and we could have periods of rain through 1 pm Saturday. We may pick up 1-2 additional inches of rain here.


Storm Day! Part 2

Greetings from rain-soaked Bodega Head!

We have now picked up over an inch and a half of rain since this morning at 7 am. We have had periods of light rain, heavy rain, big ole fat rain, and rain that even seemed to come up from underneath. The short range, high-res MM5 forecast suggested that we would get 1.5 inches for all of wednesday:


24 hr accumulated precipitation forecast by 4 km MM5 valid at 4 am UTC on Feb 26, 2014. Credit: DRI-CEFA.

We reached Atmospheric River conditions here at the BML HMT station just after midnight and have been under an AR ever since. The surface met time-series over the last 48 hours or so looks like:


Sfc met variables collected at the 10m tower at BML for the 48 hours ending at 5pm PST on Feb. 26, 2014. Remeber that the time ordinate increases to the right! Credit: NOAA-HMT.

My favorite panels are the lowermost 2. Winds have been southerly or south-southeasterly all day. This has been the case throughout deep layer overhead:


Time-vertical profile of low-layer winds, freezing level, and upslope water vapor flux. Credit: NOAA-HMT.

Since the 449 MHz wind profiler is alive and kicking again, I couldnt resist posting a plot of the near-surface winds. For the most part the wind profiles themselves have been uniform and southerly in the lowest 5 km. The approaching storm is a very deep system, and since we are on the east flank, we will continue to get southerly winds until the first cold front passes us.

The radar senses virtual temperature (via density) using RASS. The black dots in the later stages of the picture correspond to the freezing level above BBY, which is currently about 1.9 km high. Notice that after 2 pm, the winds in the lowest 2 km have veered. As a result, the freezing level has raised slightly. Veering winds correspond to warm-air advection (WAA), so this low-level warming should be expected.

We did not have a prolonged period of westerly winds as this storm approached, and therefore did not see a coastal jet. Interestingly, even though we did not, and we have not had terribly stable stratification, Cazadero has recieved less rain this event than BBY. Upslope water vapor flux corresponds to the low-layer integrated component of the wind which points from BBY to CZC multiplied by the water vapor mixing ratio. In the lower panel of the plot above, this value has been increasing along with wind speed. It is expected that as upslope wv flux increases, orographic enhancement will cause more rain to fall at CZC.

CZC has thus-far recieved near 1.4 inches. Often, this rain has formed in a deep layer extending above the freezing level:


Timeseries of S-Band radar signal-to-noise ratio above CZC. Credit: NOAA-HMT.

The freezing level can be diagnosed from this plot by the location of the “bright band” of echos near 1.9 km.

Even though we have struck out on it so far, we may yet get a short lived coastal jet. The first front is expected to come ashore at BML tonight near 10 pm. This is depicted in the MM5 forecast below:


925 hPa winds (barbs and isotachs) from 4 km MM5 valid at 0600 UTC Feb 26, 2014. Credit: DRI-CEFA.

After the front, we expect a short (until 1 pm, PST Thursday) period of westerlies, during which we may have the right conditions for a jet.

Winds will turn southerly again Thursday into Friday, until another front reaches us in advance of Friday’s storm system.

At this point, to call today’s and Friday’s system separate is probably improper. The upper level potential vorticity maximum associated with Fridays storm is so rapidly deepening, that it will engulf the PV max associated with today’s storm soon:

See that here Credit: U Hawaii

At the lower levels, this rapid upper level deepening is going to intensify a small, but intense extratropical cyclone which will hit California early Saturday morning:

See the fun cyclogenesis Credit: U Hawaii

Just for fun, I thought I would post what the high-resolution forecast of this event looks like at 72 hours simulation time.


10 m wind and SLP from 4 km MM5 valid at 12 UTC on Mar 1, 2014. Credit: DRI-CEFA.

Is that a west coast TC?

Other curiosities from today:

In this post, I mentioned that we encountered low level dust this morning, and that this system was relatively very dust laden. We also found Petaluma Gap flow for most of the morning. This can be seen in the surface station map from near 10 am PST:


Surface station temperature and sea-level pressure from stations in N. California. Credit: Meso-West, U Utah.

High pressure over the CV and East Bay, compared to relative low pressure in the North Bay and coastal Sonoma, are causing flow through the Petaluma Gap to our site in this picture.

As a result, we measured very elevated particle counts, NO, NOx and Ozone for much of today. It’s a strange result, but sometimes when it is raining the air is more polluted at Bodega Bay than when it is sunny and high pressure is in control.

This is especially true in light of the data we have recently worked up for periods of quiet conditions from early in the trip. We now have an idea of just how clean the air can be here when flow is onshore. A post reviewing that information will appear here tomorrow.


which is going to accelerate and intensify Friday’s system, so that it will come right on the heels of today’s rain.

From Mongolia, with Love

Greetings from a rain-soaked Bodega Head!

We have picked up nearly an inch of rain since 7 am PST.

I missed a post last night due to an internet outage, so I’m going to leave two posts here this afternoon. Ill get to the weather, our rain outlook for the rest of the day, Petaluma Gap flow and maybe a Coastal Jet in post two. For this post, I want to revisit the cross-Pacific transported dust we have been anticipating.

Yesterday evening’s GEOS-5 run was still bullish on our chances of seeing dust pass overhead this morning near 10 am:


Dust AOT over North America valid at 1800UTC on Feb 26, 2014. Credit: NASA-GMAO.

Lo and behold, we started picking up many dust-like spectra from our ATOFMS, Laverne, around mid-morning. We also saw ice-nuclei activity in the CFDC increase. If you animate the following loop:

GEOS-5 dust global loop Credit: NASA-GMAO.

You can watch the nature of this dust as it enters the eastern Pacific. Dust travels around the south flank of the storm as it heads East, and in doing so, eventually intercepts the Pacific coast. Dust traveling this pathway certainly would have had to cross the baroclinic zone associated with a trailing cold front from West to East, meaning the air parcels would have experienced forced descent along isentropic surfaces. Thus, if the forecast was reality, any dust traveling around the South flank of the storm would have descended toward the surface.

Our instruments here at the surface are seeing dust, so score one for the model.

We are also expecting that dust may still be present at high levels after the cold front with this storm. That feature is not well captured in the model forecast, so why are we anticipating it?

The storm became occluded on Monday, Feb. 24 while it was still in the Central Pacific. Here is a surface analysis for 1800 UTC on 02/24:


Surface analysis for Gameboy Color (R) valid at 1800 UTC on Feb 24, 2014. Credit: NOAA-OPC.

During this time, dust from a large plume leaving NE Asia was making it’s way across the Pacific Ocean too. Before we try to find it in relation to the storm, lets look at the MODIS Level 2 cloud top temperature for this approximate time:


Cloud top temperature over the Central Pacific on Feb 24. Credit: NASA-GSFC.

The MODIS (Terra) swath over the storm is from approximately 20 UTC. The deep purple colors indicating very cold cloud tops form a comma shape, which is classically associated with an occluded extratropical cyclone. The left inside of the comma is near the low pressure center and is an area where dry air is intruding into the storm circulation. It is relatively free of clouds.

MODIS-Terra observed aerosol optical depth in this relatively cloud free region (the aerosol products rely on visible wavelengths and cannot function over cloudy regions.):


Aerosol Optical Thickness at 550 nm over the North Central Pacific on Feb 24, 2014. Credit: NASA-GSFC.

The cloud free region shows elevated aerosol optical thickness, suggesting that the dust has made it into this  part of the storm circulation, which is behind the cold front, along with the dry air.

For us, this means we are expecting more dust to be included in the rain samples and ambient air near the time the cold front passes. Observing the shift in insoluble residue chemistry from the rain samples may tell us something interesting about the chemical history of the dust as it was processed during its journey.


Mid-February Wind Review

Greetings from Bodega Head! Today will be a short post. We are still anticipating quite a storm to arrive later in the week, and today should be our last day in the bright sunshine, northerly afternoon winds and relatively high pressure until the beginning of next week.

In anticipation, some of the crew got out for some R&R yesterday afternoon. After so much time in the coastal grassland, its easy to forget that whole communities of giants live just minutes away:


BBACPAX at the Armstrong Redwoods State Nature Reserve.

Local Conditions:


Timeseries of local weather conditions at the Bodega Marine Reserve. Credit: DRI-WRCC.

We have had southeasterly winds for most of the morning, leading to elevated accumulation mode particle counts at the trailer (3000 particles per cc) and few course mode particles. Unlike over the weekend, we did not see any NOx, or CO spikes from the campground to our East.  We are nearing the common time for the sea-breeze to come onshore. This has happened between 10 am and 11 am PST every morning during episodes of high pressure, and we are preparing to sample elevated coarse mode (primarily sea salt) aerosols throughout this afternoon. This may be the last day we see a diurnal shift in ambient aerosol conditions for quite some time. The animation I have linked below shows the high-resolution MM5 forecast 10 m winds from this morning through late Tuesday.

MM5 surface wind loop Credit: DRI-CEFA.

As time passes, the elongated clockwise wind pattern offshore dissipates in favor of strengthening southerlies at the coast, followed by strong southerlies throughout the domain and eventually, it is possible that we get a coastal jet windward of the Sonoma County coastal range. There are also some slight hints that we may see Petaluma Gap Flow – something we will be watching with great interest.

Review of wind conditions for the first half of our study:

We are approaching the halfway point of BBACPAX, and with a prolonged period of stormy weather approaching I thought it prudent to review the primary flow regimes of the past 13 days here at the site.

In this post, I presented an analysis of the typical quiescent day here at BML. The flow we were seeing at our sampling site was dominated by a diurnal cycle related to the large-scale northerlies plus the sea-breeze circulation. For the most part, we saw strong (20 kts or more) Northwesterlies in the late afternoon hours, light south-southeast winds in the early morning, and transition periods in-between.

A wind-rose diagram of the data from the nearby DRI-WRCC station shows these preferred flow regimes very clearly:


Wind Rose generated from 10-m tower anemometer at the DRI-WRCC Bodega Marine Reserve site. Credit: WRI-WRCC.

This data has been compiled from the period Feb 12 – Feb 24, 2014. This period included one 24-30 hour period of large-scale southerlies associated with a landfalling pacific wintertime cyclone (PWC), but for the most part it was dominated by high pressure and quiescent conditions.

It is interesting to compare the Bodega Marine Reserve wind rose to that from an inland site:


The wind data in this rose comes from the Blue Oak Ranch Reserve site of the same network. The site is located windward of the Central CA coastal range but inland. Instead of two primary flow regimes here, we see simply one prevailing wind flow direction, WSW.

The diurnal flow reversal we experienced during our first two weeks at Bodega Marine Lab had a profound effect on the aerosol number, size and chemistry we observed. That data is being analyzed now, and I hope to include some relevant blog posts about aerosol and trace gas concentration (with cool figures!) soon.

Next chance for rain:

We are still watching the storm which should arrive Wednesday. Ill post tomorrow about our chances to see unique flow features such as the Coastal Jet and Petaluma Gap Flow.


Previewing the Late-Week Storm

Greetings from Bodega Head!

Local Conditions:

We woke up to shallow marine stratocumulus this morning, the first time we saw something other than sunny skies since Tuesday. We currently have lightly offshore flow, which is bringing air from the Westside Park Campground over our instrument trailer. The weekenders burning campfires there have given us some fine mode particulates and reactive gases to sample. Currently we are measuring:

1500 particles per cubic centimeter (cc) on the CPC

An accumulation size mode near 65 nanometers (nm) on the SMPS

A nearly flat APS trace, with total “course mode” particle count of 6 per cc.

Elevated NOx at 7 parts per billion (ppb)

Elevated CO at 180 ppb

These numbers are a dramatic shift from our typical afternoon measurements. Under the influence of strong onshore flow, we typically see NOx near 0 ppb, CO near 140 ppb, around 500 total particles per cc, no accumulation mode and a course mode near 800 nm with total course mode count from 12 – 15 particles per cc.

We are expecting to encounter similar conditions after the winds come onshore again this afternoon.

Medium Range Forecast Drift:

For the last few posts I have been pointing to the end of this week as our next chance of rain. The models were converging (GFS and ECMWF agreed) on a storm hitting California on Friday 2/28 and Saturday 3/1. This is still the favored solution in the models. In fact, this storm has only gotten more robust given new forecast initial conditions. The tendency of a solution at a particular time to change given a change in initial model conditions is referred to as model drift. You can think of it like this:

I am interested in a storm on Friday, February 28 at 1200 UTC.

The first hint of this storm came in the 240 hr forecast from the ECMWF model initialized at 12 UTC on February 18. The hint looked like:


500 hPa geopotential height over the Northern Hemisphere from the ECMWF Ensemble Prediction System. Initialized at 12 UTC on Feb 18. Valid at 12 UTC on Feb 28. Credit: ECMWF.

In this scenario, February 28 at 12 UTC is the forecast “valid time” and February 18 at 12 UTC is the forecast “initial time”.

If I keep the valid time the same, but look at the model forecast intialized on the next day, I will start to get an idea of how updated information about the state of the real atmosphere on February 19 changed the forecast for February 28:


Same as last figure,  but initialized at 12 UTC on February 19.

Model drift can be a qualitative way to examine whether new information adds confidence that the forecast event (our storm) will occur or whether this new information makes the event forecast less confident.

Lets go through the progression or new initial times for our Friday storm, ending with yesterdays initial time:


Same as last figure,  but initialized at 12 UTC on February 20.


Same as last figure,  but initialized at 12 UTC on February 21.


Same as last figure,  but initialized at 12 UTC on February 22.

We see that after each new day’s initialization, the upper level trough nearing the West Coast on Feb 28 becomes deeper, until on the last intital day (Feb 22) there is a dramatic deepening, and the upper level feature in the ensemble forecast now looks like a storm in a deterministic model, i.e. it looks like a real world event.

This is good news for our storm. Over time, the ensemble of simulations became more confident that the event (a strong upper level trough, supporting a winter storm on Friday, February 28) would occur.

There is something else of note that has happened to our storm, however.

To see it, examine this animation of upper level wind flow forecast by the GFS:

GFS 600 hPa loop Credit: U Hawaii.

Start at the valid time 18 UTC on February 24 (30 hours into the forecast). There is a closed upper level low in the geopotential height field near 4oN; 145 W. Stepping through the forecast from here, we can see this closed low “retrograde” back toward the Aleutian Islands. As the low moves out of the Jet Stream (the belt of high wind speeds), two shortwave eddies in the height field move into its place. As these rotate counter-clockwise around the larger scale trough, they intensify. The first spins up a storm which should hit our area late on Wednesday, Feb 26. The second becomes a deep trough, then a cut-off low which becomes our Friday storm. We may be in for a prolonged period of wet weather this Wednesday through Saturday!

Other storm characteristics:

Our partners from CalWater-2 just finished a few exploratory flights into Pacific atmospheric rivers (AR). AR commonly reach California in association with storms like the one on tap for later in the week. If the broad scale circulation can pick up enough water vapor from the tropics and advect it Northeast, a filament of integrated water vapor (IWV) or precipitable water (PW) is visible in satellite imagery. We can look at a model mock-up of the same metrics courtesy of U Hawaii:


Tropospheric Precipitable Water from GFS valid at 00 UTC on Feb 27, 2014. Credit: U Hawaii.

The first image is PW over the Pacific ocean on Wednesday afternoon (4 pm PST). There is a tongue of elevated moisture which is expected to hit the central California coast in concert with the arrival of the Wednesday storm. This filamentary object in the PW field is not as impressive as a classic AR, but it is good news for the prospect for rain.

For Friday:


Same as above figure, valid at 12 UTC on Feb 28, 2014. Credit: U Hawaii.

For Friday, we see a plume of elevated moisture also arriving along with the second storm. This happens early in the morning on Feb 28. As before, the maximum PW content is not quite up to AR standards, but it is better than none.

Long-Range Transported Dust:

Our BBACPAX crew is anxious to observe some dust particles and measure their ice nucleating abilities. We may get our wish during the next two storms. GEOS-5 predicts a very large dust plume to leave the Asian continent today, travel across the Pacific near 30 N and wrap around the circulation of our developing storm systems by sometime Tuesday. This mid-Pacific dust belt will continue to feed into the storm’s southern flank through Friday, which may elevated dust aerosol optical depths over California to levels we have not observed during our trip.  See the animation of the GEOS-5 dust AOT forecast at the link below:

GEOS-5 Global Dust Forecast Credit: NASA-GMAO.

The questions are: Will the dust behave as forecast, and will we observe it at the surface?

We cannot yet answer the first question. We may have validation information from satellite data tomorrow after the first A-TRAIN flights over the Plume region east of Japan. However, unlike the weak plume forecast on Feb 20 discussed in this post, the dust plume forecast by GEOS-5 in this case is massive, meaning there is less chance that the model misses some rainout, surface deposition, or flow perturbation by half a degree that may cause the plume to disappear, dissipate, or dive south before its forecast target.

For the second question: The dust in the model most likely makes the first part of its journey at higher levels where winds are fast. It must do this in order to cross nearly half the Pacific ocean in 2 days. However, on our side is the fact that it must travel from west to east through a cold front on the south side of the developing storm. In doing so, the air carrying the storm must descend in order to stay on isentropic surfaces. The fact that there is a new baroclinic zone associated with the wednesday storm means that air parcels which follow the southern pathway will likely have to descend twice before crossing the coast. This improves our odds at seeing some dust at the surface.

Working against us is that during these descents, the dusty air will also likely encounter a lot of rain, which will scavenge most of it from the atmosphere. However, this is precisely the phenomena we want to catch using our rain samplers: aerosol interaction with cloud microphysics.


If the three-phases of our developing storms: dynamics, water vapor, and long-range dust line up as the models are predicting, this could be an epic event both for our scientific goals and for the badly strained water resources of California. This blog will be keeping track of the developing storm for the rest of the week.

Cross-Pacific Dust

Greetings from Bodega Head!


Its a pleasant sunny day with light WNW winds. We are still under the control of the NE Pacific ridge and as a consequence, we are stuck with miserably nice weather.

Yesterdays conditions here at Bodega Marine Lab were considerably less windy than in days past, due to slight offshore flow dominating the local scale. See the figure below for a reference point:


West Coast surface analysis for 18 UTC on Feb 20, 2014. Credit: NOAA-OPC

This flow pattern took some of the bite out of the sea-breeze. The timeseries of our nearby anemometer  shows the calm conditions, but with a shift to offshore flow during the overnight hours (i.e. during the sea-breeze reversal).


48 hour timeseries of surface weather variables at BBY. Credit: NOAA-HMT.

There isnt much else to report in local weather news. We are still anticipating a storm reaching the Northern California coast next Friday and Saturday. Ill update the forecast of that storm tomorrow.

Long-Range Transported Aerosol:

Yesterday we were excited by the possibility that a small plume of enhanced dust concentration may cross our area. According to global circulation model (GCM) / chemical transport model (CTM) forecasts This plume originated in Asia several days ago, was cut off from it’s parent plume by a mid-Pacific storm, and then would have been steered clockwise around the ridge to our location. We were hoping to draw some of this dust through our sampling manifold, cause it to participate in heterogeneous immersion ice nucleation in our CFDC and then investigate the ice crystal residue chemistry through a combination of a pumped counterflow virtual impactor and an aerosol time-of-flight mass spectrometer.

The impetus for believing the dust would reach us came from a forecast made by the NASA global modeling and assimilation office (GMAO) GEOS-5.


Tropospheric dust AOT forecast valid at 3 pm PST on Feb 20, 2014. Credit: NASA GMAO.

The GEOS-5 is a global circulation model which contains some simplified aspects of a chemical transport model. Dust mass mixing ratio is forecast based on emission from a saltation-imapction type emission model and the GEOS-5 3 dimensional winds. The model shortwave emission model is used to calculate the aerosol optical thickness (AOT) which would result from the dust mass and vertical distribution in each atmospheric column.

Unfortunately, we did not detect elevated dust levels at the surface yesterday afternoon in BBY, meaning that the point forecast for our area of the GEOS-5 domain may have busted. The caveat here is that we are only privy to the total column dust AOT through the web portal. It could be that dust in this plume simply never reached the surface.

Let’s examine further cross-pacific dust transport and how we might verify global model (such as GEOS-5) forecasts.

Recall that in a previous post I posted an image from MODIS-Aqua which indicated a region of high total aerosol optical thickness leaving the Asian continent in a large plume. That image was a composite of all “swaths” (the footprint of a satellite instrument) of MODIS level 2 aerosol optical thickness (AOT) at 550 nm from February 17. Here is the corresponding image from 02-18:


North Pacific swaths of MODIS level 2 AOT from Feb 18, 2014. Credit NASA-GSFC.

There is still a large area (the size of the entire western US) of elevated AOT east of Japan. The satellite carrying MODIS, Aqua, crossed this aerosol plume near 16:30 UTC. Lets see what the GEOS-5 dust forecast looked like for this time and date:


Tropospheric dust AOT forecast valid at 17 UTC on Feb 18, 2014. Credit: NASA GMAO.

The model does predict a plume of moderately optically thick dust extending east from Asia over Japan at the correct date and time. For now, we will not worry about whether the model gets the magnitude of AOT correct, but rather simply worry about a “qualitative plume” arriving at the correct location in space and time. We should note that the meridional extent of the plume in GEOS-5 is smaller than it appears in MODIS AOT.

The next question we should ask is “Is the dust confined to a discrete vertical layer?” followed by “If so, how elevated is the dust layer?” To partially answer this question we can turn to CALIOP. CALIOP is an active, polarized, visible light emitting/receiving interferometer. It flies on CALIPSO in the same orbit and slightly behind Aqua. When the atmosphere is optically thin (no thick clouds) CALIOP can estimate the vertical profile of aerosol optical extinction (related to AOT) at nadir. Lets look at what CALIOP saw along this path:


One orbital path of the CALIPSO satellite on Feb 18, 2014. Credit: NASA-LARC.

The pink section of the above swath passed over the approximate area of our plume. The measurement along the pink section looked like:


532 nm attenuated backscatter (top) and target classification (bottom) from CALIOP along a path which crossed the northeast Pacific on Feb 18, 2014.

The top plot is the time-altitude cross section of 532 nm attenuated backscatter. The plume we are interested can be seen between 134.5 and 139 E longitude as a shallow near-surface layer of yellow shading. The bottom plot is generated based on backscatter and the 532 nm depolarization ratio. After data is collected, regions of significant backscatter signals are classified based on their likely size and shape relative to Mie spheres. According to the classification algorithm, this near-surface layer is full of aerosol.

So far, so good. We identified a layer of aerosol in MODIS data, its timing and position was forecast by GEOS-5, we saw an aerosol layer with CALIOP, and are now confident that this aerosol layer is confined to the lowest 3 km.

What about the aerosol plume that was supposed to cross the California coast yesterday? Unfortunately, the data from the CALIPSO swath which passed over this area is not yet available to view online. However we can check MODIS:


North Pacific swaths of MODIS level 2 AOT from Feb 20, 2014. Credit NASA-GSFC.

MODIS identified very low AOT in the cloud-free pixels near the north-central California coast during its overpass on Feb 20, 2014. This area was crossed by MODIS near 2230 UTC. It is near the same area forecast by GEOS-5 to contain elevated dust aerosol optical depth at 2300 UTC (see third figure, this post).

In this case, the GEOS-5 model and MODIS imagery
do not agree on the timing and location of an aerosol (possibly dust) plume. Our instruments also did not encounter dust spectra at a rate above background level, so it may be that the GEOS-5 forecast lead us on a goose chase yesterday.

Going forward, we will continue to use global model dust forecasts to guide our measurement strategies, but they must be taken with a large remote-sensing grain of salt.

Tomorrow’s post: Medium range convergence of global model forecasts.