Satellite imagery to show strategic information has been national space programmes primary function from deep within the Cold War.  The US ‘Discoverer’ series (CORONA spy satellites hidden behind a civilian name) were the first western Earth-observing systems launched, years in advance of far cruder meteorological imagers (1960) and more than a decade before environmental monitoring started with Landsat-1 (1972).

So what do modern systems say about timeliness of monitoring modern strategic information?  What’s interesting about the conflict in Syria is the speed of online delivery systems has a by-product that some of the large-scale activity on the surface are visible in near-real-time to casual users.  No dedicated, nor top secret, monitoring system is necessary.

The NASA Terra MODIS imagery from the early pm overpass on 23rd November shows two conflagrations to the SW of Raqqa, the stronghold of ISIS in northern Syria.  The pink overlay is a shapefile of the Syrian road network, which are free to download.

Enter a caption

These dark clouds of smoke are are most likely from some form of oils storage facilities burning. These could be tankers or larger, static oil tanks.  The spatial resolution of MODIS is 250 m, so there’s a lot of smoke with these- the most northerly plume stretches for tens of kilometres.  Static containers are most likely

Here’s the MODIS image above, bit as a close up of the fire closest to Raqqa.


Well, what can we tell?  The fire is located near a road, but not on one, so it’s infrastructure, rather than transport that’s burning.

How can we investigate further?  We need greater spatial resolution.  The next (free, unclassified) stage is Landsat.  As discussed previously, Snapsat have produced the most user-friendly interface imaginable.  Even better, the downloads are now in geotiff format- you can slot them straight into your favourite GIS tool. Landsat has a 17-day revisit cycle. Luckily, there was a Landsat 8 overpass the next day and, most importantly, it was cloud-free.

This is an ‘urban’ 7-6-4 mix, but many other combinations available.


There’s a group of oils storage containers, just from where the smoke is emanating. It appears that only one of the storage cylinders exploded, located in the SW corner of the complex.  Here, a day later, there’s still a faint trail of smoke visible. It’s the faint smudge, seen drifting NW.

We can see this in greater detail using the USGS’s Explorer site to add the 15 m spatial resolution band 8.


This is a clearer product spatially, but unlike the multispectral Snapsat product, there’s no atmospheric corrections- it’s raw.  This gives a washed out feel.  Doing corrections isn’t too hard, but for quick-looks like this, it’s time-consuming. Also the reflectance products from USGS are currently suspended (starting 01/11/15) due to issues with TIR, so that removes another layer of clarity.

A 1:10 000 scale image highlights the superior spatial resolution that to a skilled analyst, or just someone who knows the area, can provide immense information.  It could also be that the fire is out and those grey streaks are just an extension of what seems to be a Wadi to the south.


The other source of near-real-time visible data is NOAAView- but that’s available one day later (shows the previous day) and is only 750 m spatial resolution – these sort of applications were secondary  or even tertiary considerations in its design.  Here is 23rd November mid-afternoon.  The clouds of dark, oily smoke are dissipating.


However, there’s still some useful information.  In this later afternoon overpass, the fires are still burning, but no new activity can be seen.

What does this set of images tell us?  Lots, if you’re an interested party. But at their base-level, they illustrate that high-quality strategic information is available to everyone, if you know where to look for it. Also, they’re easy to obtain: it took longer to write this blog post than to download and process all the images.


Lake Chad

Sitting down over breakfast with an old Readers Digest World atlas (you may have one!), my son was perusing the facts and figures at the back. “Lake Chad, where’s that?”  It’s one of the largest lakes in Africa, says our 1960’s reference.  To the internet to see what it’s like now and we get an NASA Earth Observatory page showing the lake in 2001. This shows how much it’s dried up – 95% gone- since our map was printed.

My son asks, “Has it now vanished completely?”  To Nasa Worldview.  The first almost cloud-free image is just a few days ago, from June 29th, and yes, it looks like there’s nothing there!  Sand where water should be!  Hold on, is it just a murky lake?  Let’s change channels from VIS to solar IR (reflected from the sun, not emitted by Earth’s heat), and yes, there’s still water.

The IR channels on Modis are only 500 m (250 m for the visible), I wonder if Landsat with 30 m resolution can help.  Off to and scan across to Path 185, Row 51 and find an image with 1% cloud from less than a fortnight ago.

It’s a 5-4-3 ‘vegetation’ combination of channels, producing false colour where red is the health and depth of plant life- the more red, the more green in reality.  The marshy surface of the former Lake Chad really stand out. But on our screen it’s also a quite beautiful scene.

Landsat Lake Chad
Landsat 564 185/051 21/06/2015

This is only the preview image- click on ‘Download a full-size render’ at this link for the 30 m version.

That helps my son’s question, but other Landsat channel combinations help us with the land-water discrimination.

5-6-4 provides some interesting detail.

Landsat Lake Chad
Landsat 564 185/051 21/06/2015

A screenshot of the lower-left-hand corner of the extant lake shows some exquisite detail.

Lake Chad detail 21/06/2015 5-6-4
Lake Chad detail 21/06/2015 5-6-4

The blue seems to be a function of water depth (not being able to see the lake bottom?), but perhaps it’s actually clarity?  Looking at this image, you’d assume that the former lake-bed would be marshy- yet the bright green line is an active fire, the smoke drifting northwards above it underlining this.  Slash and burn is hardly new in agriculture, but one wonders how long this part of a lake-bed was ever dry enough to be flammable?

I’ve not started looking to carefully at Lake Chad, this is only as a taster.  There’s SO much going on, it’s almost like Landsat was designed to view this type of area. This post is another illustration of the immediate nature of available data.  Choosing, processing and downloading all the images took less than 10 minutes in total.  The availability of this much information, so readily processed, is a modern miracle.  However, like so, so much satellite data, it raises as many new questions as it answers.  I think that’s a good thing as it keeps me coming back.


Path length

School physics classes talk about scattering of light and how it makes the the sky blue (but why not violet?) and sunsets are red (although they are different- another post sometime).  The  controls on the amount of scattering effective are the number of particulates and the path length the light has to travel through the atmosphere.

Path length is simply sun angle to nadir (see the related concept of ‘air mass’, which is confusing in this context). If the sun is sat vertically above you, that’s one atmospheric path length.  By the time it’s on the horizon (technically it’s below the horizon due to refraction), the path length increases to 20-40 (literature varies on the estimate, would you believe).

A great illustration of this can be seen in this morning’s Meteosat imagery from Sat24.  As dawn was breaking at 0400 UT, the UK’s east coast appear to have an anomalous cloud feature:

Meteosat visible imagery 0400 UT 01/07/15
Meteosat visible imagery 0400 UT 01/07/15

An hour onwards, we see the ‘cloud’ feature ‘weakening’ , then fading by 0805 UT,

Meteosat visible imagery 0500 UT 01/07/15
Meteosat visible imagery 0500 UT 01/07/15
Meteosat visible imagery 0805 UT 01/07/15
Meteosat visible imagery 0805 UT 01/07/15
Meteosat visible imagery 09050 UT 01/07/15
Meteosat visible imagery 09050 UT 01/07/15

and by 1105 UT almost invisible.






So what causes this reducing brightness?  Quite straight-forwardly, that as the atmospheric path length of the sun decreases, scattering by particles (Rayleigh) also decreases.  The feature (if it has a name?) will weaken and become slowly less apparent.

I wonder if there are any sunrise pictures taken this morning from the English east coast?  It would be a big, red ol’ sol.

The next question is why are we seeing it now?  Well there must be a large amount of large particles- this highly likely due to many factors.  First, the prevailing wind in the planetary boundary layer (the bit of atmosphere next to the surface) was south-easterly- see this US GFS model of winds (via this lovely presentation).

GFS surface winds,  TPW 1000 UTC 01/07/15

This brings air from the continent with anthropogenic pollution and  natural dust particulates from land.

Also, the air was very humid. The next images show the SSMIS passive microwave image for 0515 UT and the 0500 IR Air Mass product from MSG.  Total Precipitable Water is many lower tropospheric humidity and, away from the coastal mask,  you can just see it falling away towards the middle of the southern North Sea.  It seemed to have reached about 30 kg m-2, which was very high for the UK. The IR is a bit messier with little quantitative value, but it certainly suggests there’s an east-west gradient over the North Sea, which was probably humidity based.

MSG airmass
MSG Airmass product 0500UT, 01/07/15
SSMIS TPW 0515 UT 01/07/15

Many dust and pollution particulates, like sulphur dioxide, are hygroscopic.  They love water vapour and grow in size considerably if the conditions were right.  It’s not really my area, but I guess that the conditions were right!

So water vapour-loving dust particulates, high humidity and a low sun angle produce some very funny effects in the imagery- one to watch out for this evening.  And yes, as the sun sets in the west and the path length increases…


Meteosat visible imagery 1745 UT 01/07/15
Meteosat visible imagery 1745 UT 01/07/15
MSG visible
Meteosat visible imagery 1915 UT 01/07/15

A novel space thermometer

A first post using data from an old friend- NASA’s Modis instruments.  True-colour imagery at 250 m resolution is available in a  number of formats from NASA Worldview.

So today, something caught my eye- can you see it…? It’s just north of Hull in NE England

image-download (1)

No?  Perhaps a little closer


An elongated white line; white as it is saturating all three colour channels.  Looking back over the last few weeks (and it took a few weeks to find a cloud-free image), there’s nothing like this visible.  So what is it and why today?

Worldview allows you to export the imagery into a Google Earth kmz file- this provides a quick and dirty GIS reference.  Opening up the area in GE,  we find a bank of commercial glasshouses at Thearne- you can see them on a map.

The title of the post is ‘a novel space thermometer’, so how does this large reflection from greenhouses help with the temperature?  It is assumed that most of the time, fruit and vegetables are needed to be kept warm, so the glass roofs will be shut.  Today though it’s 27 C in that area and it’s the hottest day of the year across much of England.  The hundreds of glass panels were therefore moved into an open position to allow heat out.  This position just happens to reflect the light perfectly into the orbiting Modis sensor, producing a surrogate ‘hot’ reading from space.

There are a number of commercial greenhouses unexpectedly flashing at the satellite today- Smallford near Hatfield in Herts., a particularly well-aimed reflection from a small structure in Stotfold in Beds., and points before and after. Notice how they’re all in a line.  This demarcates the area directly below the satellite, its ground track.

Theories are only as good as the predictions they make.  Tomorrow, the glass will be in the same position as it’s the forecast is for an even hotter day – but with risk of thunderstorms- what happened to the three hot days before that we’re normally promised? 🙂  –  however, the Modis orbital track will be to the side, not overhead.  So if the weather is as hot and sunny on Thursday, we should see them again.