WEBVTT
Kind: captions
Language: en

1
00:00:00.170 --> 00:00:01.820
- This week's guest video is from a channel

2
00:00:01.820 --> 00:00:05.140
based around explaining
interesting things, on location,

3
00:00:05.140 --> 00:00:08.720
with a style that sometimes feels like
a BBC documentary from the early 2000s.

4
00:00:08.720 --> 00:00:12.040
So, introducing someone who is
going to have my job in a few years,

5
00:00:12.040 --> 00:00:13.840
James, from Atomic Frontier,

6
00:00:13.840 --> 00:00:16.390
who starts his video about
bushfires and satellites

7
00:00:16.390 --> 00:00:17.960
by climbing a tree.

8
00:00:17.960 --> 00:00:19.453
James, take it away.

9
00:00:20.480 --> 00:00:23.480
- I'm up a tree in the land of Down Under.

10
00:00:23.480 --> 00:00:27.970
Not only does this perilous position
give us a convenient place to hide

11
00:00:27.970 --> 00:00:30.910
from the hordes of killer
kangaroos down on ground level,

12
00:00:30.910 --> 00:00:34.140
but it also provides a
convenient vantage point

13
00:00:34.140 --> 00:00:37.430
from which to look for another
very Australian threat.

14
00:00:37.430 --> 00:00:41.550
Since the 1930s, lookout trees
have been used to spot bushfires

15
00:00:41.550 --> 00:00:44.460
in the giant karri forests
of Western Australia.

16
00:00:44.460 --> 00:00:48.540
This one, the 65 metre tall,
Dave Evans Bicentennial Tree,

17
00:00:48.540 --> 00:00:51.080
is the tallest tree house in the world.

18
00:00:51.080 --> 00:00:55.400
From the top we can see up to
20 kilometres in every direction.

19
00:00:55.400 --> 00:00:58.590
However, to monitor something,
the size of a continent,

20
00:00:58.590 --> 00:01:00.740
we're going to have to climb even higher,

21
00:01:00.740 --> 00:01:03.630
to the satellites orbiting
far above our heads.

22
00:01:08.330 --> 00:01:10.050
Here is a series of images

23
00:01:10.050 --> 00:01:12.800
taken by Japan's Himawari
weather satellite.

24
00:01:12.800 --> 00:01:17.090
It's in a geostationary orbit where the
24-hour rotation of the Earth

25
00:01:17.090 --> 00:01:20.660
is exactly matched by the
satellite's own 24-hour orbit.

26
00:01:20.660 --> 00:01:25.060
Such orbits are only possible
at an altitude of 36,000km

27
00:01:25.060 --> 00:01:27.070
and directly above the equator.

28
00:01:27.070 --> 00:01:29.610
Himawari takes a photograph
every 10 minutes

29
00:01:29.610 --> 00:01:34.010
giving meteorologists the ability to
observe weather as it develops in real time.

30
00:01:35.000 --> 00:01:38.860
Between June 2019, and May 2020,

31
00:01:38.860 --> 00:01:42.250
18 million hectares of
Australian bush burned

32
00:01:42.250 --> 00:01:44.950
in what is now known as
the Black Summer.

33
00:01:44.950 --> 00:01:49.090
34 people, and more than 4 billion
terrestrial vertebrates

34
00:01:49.090 --> 00:01:50.840
perished in the flames.

35
00:01:50.840 --> 00:01:54.780
From its unique vantage point
the Himawari weather satellite watched on

36
00:01:54.780 --> 00:01:59.060
as cities on the eastern seaboard
were smothered in the smoke.

37
00:01:59.060 --> 00:02:02.580
Geostationary satellites aren't
a great fire detection tool.

38
00:02:02.580 --> 00:02:06.250
By the time the smoke is visible
the flames beneath have grown too large

39
00:02:06.250 --> 00:02:07.620
to control effectively.

40
00:02:07.620 --> 00:02:11.430
We spotted the fire, but your
house burned down six hours ago.

41
00:02:11.430 --> 00:02:14.630
By dropping from a
geostationary to low earth orbit

42
00:02:14.630 --> 00:02:17.000
we can massively improve resolution.

43
00:02:17.000 --> 00:02:19.910
However, this comes at the cost
of no longer being able to match

44
00:02:19.910 --> 00:02:21.990
the rotational period of the Earth.

45
00:02:21.990 --> 00:02:25.450
To achieve live coverage we therefore
need a lot of satellites,

46
00:02:25.450 --> 00:02:28.750
so that as one leaves an area,
the next one arrives.

47
00:02:28.750 --> 00:02:31.510
Most fire services can't
afford to buy multiple

48
00:02:31.510 --> 00:02:34.720
290 million-dollar apiece
weather satellites,

49
00:02:34.720 --> 00:02:38.320
but these small cubesats
come in at a fraction of that cost.

50
00:02:38.320 --> 00:02:40.290
You can launch about 4,000 of them

51
00:02:40.290 --> 00:02:42.820
for the same price as
one weather satellite.

52
00:02:42.820 --> 00:02:46.300
CubeSats were originally
intended for low-cost research,

53
00:02:46.300 --> 00:02:49.740
but after more than 15 years
of research and development

54
00:02:49.740 --> 00:02:52.070
they're now good enough
for use in industry.

55
00:02:52.070 --> 00:02:55.840
This one, built by Curtin University
in Western Australia

56
00:02:55.840 --> 00:03:00.520
will eventually be able to take photos with
a resolution down to 4m across.

57
00:03:00.520 --> 00:03:02.510
Rather than just looking for smoke,

58
00:03:02.510 --> 00:03:05.620
at this resolution we can
also spot the fire itself.

59
00:03:06.610 --> 00:03:09.960
Flames are red, so by
including a simple red filter

60
00:03:09.960 --> 00:03:12.630
we can quickly zero in on the target.

61
00:03:12.630 --> 00:03:15.270
Unfortunately, as you can see
from my assorted collection

62
00:03:15.270 --> 00:03:19.970
of car boot sale items, which yes,
does include the obligatory red t-shirt,

63
00:03:19.970 --> 00:03:22.080
we're likely to get a lot
of false positives

64
00:03:22.080 --> 00:03:25.580
from things that are red,
but also not on fire.

65
00:03:25.580 --> 00:03:29.720
To refine the search further we
need to include another condition.

66
00:03:29.720 --> 00:03:31.680
Fires have a characteristic flicker

67
00:03:31.680 --> 00:03:34.980
growing brighter and dimmer
roughly 12 times per second.

68
00:03:34.980 --> 00:03:36.870
By first searching for the colour red,

69
00:03:36.870 --> 00:03:39.630
and then checking that it
flickers at this 12 hertz,

70
00:03:39.630 --> 00:03:41.580
we can detect fires when they begin.

71
00:03:42.500 --> 00:03:46.560
As the fires grow, so do their
accompanying clouds of smoke.

72
00:03:46.560 --> 00:03:49.450
After a point, we can't
even see the flames.

73
00:03:49.450 --> 00:03:52.250
This is terrible for
firefighters and residents

74
00:03:52.250 --> 00:03:55.400
who need to know where the fire is,
and its rate of growth,

75
00:03:55.400 --> 00:03:56.840
in order to stay safe.

76
00:03:57.720 --> 00:04:01.900
In order to look through the smoke,
we can use infrared radiation.

77
00:04:01.900 --> 00:04:05.720
Now, our eyes can't see in
infrared like my camera can,

78
00:04:05.720 --> 00:04:08.310
but if you hold your hand
close to your face,

79
00:04:08.310 --> 00:04:10.140
you can feel it as heat.

80
00:04:10.140 --> 00:04:14.840
Infrared radiation has a few key benefits
over the visible variety.

81
00:04:14.840 --> 00:04:17.360
First of all, there's a
lot more infrared radiation

82
00:04:17.360 --> 00:04:20.260
being produced by a
fire than visible light.

83
00:04:20.260 --> 00:04:22.620
Next, it has a much longer wavelength,

84
00:04:22.620 --> 00:04:24.600
so it's able to pass
through clouds of smoke

85
00:04:24.600 --> 00:04:26.250
pretty much unaffected.

86
00:04:26.250 --> 00:04:30.180
Together, this means that we can detect
and then track large blazes.

87
00:04:31.190 --> 00:04:33.920
Infrared cameras have
quite a low resolution,

88
00:04:33.920 --> 00:04:37.410
which means that small spot
fires will often go undetected.

89
00:04:37.410 --> 00:04:39.030
We need to monitor them, however,

90
00:04:39.030 --> 00:04:41.910
because these spot fires can
act as important precursors

91
00:04:41.910 --> 00:04:44.810
to the larger blaze,
jumping over fire breaks,

92
00:04:44.810 --> 00:04:47.200
and spreading to new patches of forest.

93
00:04:47.200 --> 00:04:50.340
Fortunately, there is one
final fire detection technique,

94
00:04:50.340 --> 00:04:51.640
which I'm going to share with you,

95
00:04:51.640 --> 00:04:54.250
especially designed for
these smaller blazes.

96
00:04:54.250 --> 00:04:56.400
However, please, don't try this at home.

97
00:04:58.590 --> 00:05:03.000
Here are three beakers, into which
I've placed about 20ml of ethanol.

98
00:05:03.000 --> 00:05:05.260
Into the first we have a pure solution,

99
00:05:05.260 --> 00:05:08.850
and the others are mix of boric acid
and potassium chloride.

100
00:05:08.850 --> 00:05:11.300
We're going to see what
happens when they ignite.

101
00:05:13.670 --> 00:05:18.030
As you can see, ethanol burns with a
clear, slightly orange-red flame,

102
00:05:18.030 --> 00:05:20.530
it's quite dim and almost invisible.

103
00:05:20.530 --> 00:05:22.890
That's because it's quite
a clean-burning fuel

104
00:05:22.890 --> 00:05:25.120
producing a minimum amount of soot.

105
00:05:25.120 --> 00:05:29.180
It's the heated soot which gives
a regular fire its orange-red flame,

106
00:05:29.180 --> 00:05:31.880
but for our experiment that
would just get in the way.

107
00:05:33.020 --> 00:05:34.550
Next, we have the boron.

108
00:05:34.550 --> 00:05:37.030
It burns with a bright green flame.

109
00:05:37.030 --> 00:05:40.090
It's not because of some special effects
or camera trickery,

110
00:05:40.090 --> 00:05:45.200
it legitimately looks like the
alchemists' wildfire in "Game of Thrones."

111
00:05:45.200 --> 00:05:48.840
The green is actually a mix
of a few very specific wavelengths,

112
00:05:48.840 --> 00:05:53.580
characteristic of the electron configuration
of the vaporised boron ions.

113
00:05:53.580 --> 00:05:54.820
When the ions are heated,

114
00:05:54.820 --> 00:05:57.640
their electrons jump up
and down the energy levels

115
00:05:57.640 --> 00:06:00.410
producing wavelengths
specific to that jump.

116
00:06:00.410 --> 00:06:03.540
Different elements have
different electron configurations

117
00:06:03.540 --> 00:06:06.460
leading to a range of possible colours.

118
00:06:06.460 --> 00:06:08.780
Trees don't contain all that much boron,

119
00:06:08.780 --> 00:06:11.620
which is why wood doesn't
burn with a green flame.

120
00:06:11.620 --> 00:06:14.120
However, they do contain
a lot of potassium

121
00:06:14.120 --> 00:06:16.440
where it's used for cell metabolism.

122
00:06:16.440 --> 00:06:19.210
The potassium burns
with a dim purple flame.

123
00:06:19.210 --> 00:06:22.030
It's a bit disappointing
in the visible spectrum,

124
00:06:22.030 --> 00:06:24.800
but it doesn't only produce
visible wavelengths,

125
00:06:24.800 --> 00:06:26.950
it also produces infrared.

126
00:06:26.950 --> 00:06:31.250
The 770 nanometer wavelength
in particular is very bright.

127
00:06:31.250 --> 00:06:33.490
Since potassium only
produces this wavelength

128
00:06:33.490 --> 00:06:36.710
during combustion, if we tune
our detectors to this band,

129
00:06:36.710 --> 00:06:39.550
we can be certain that it is
coming from an active blaze.

130
00:06:39.550 --> 00:06:43.830
With the right filters we can even
identify fires smaller than a pixel

131
00:06:43.830 --> 00:06:47.350
since the 770 nanometer
emission is so bright.

132
00:06:48.410 --> 00:06:52.290
We have come a long way from
the bushfire lookout trees of the past.

133
00:06:52.290 --> 00:06:56.880
With fleets of low-cost satellites
tuned to specific wavelengths,

134
00:06:56.880 --> 00:06:59.670
we can detect bushfires as they flare up

135
00:06:59.670 --> 00:07:04.700
and then track their progress
even as the plumes of smoke start to grow.

136
00:07:04.700 --> 00:07:08.370
The fire service may no longer
need these lookout trees,

137
00:07:08.370 --> 00:07:11.640
but, they can provide visitors
some spectacular views

138
00:07:11.640 --> 00:07:14.220
of the forests that
we're trying to protect.

139
00:07:14.220 --> 00:07:17.270
The only question is: how am
I going to get down from here?

140
00:07:17.270 --> 00:07:20.130
This has been James Dingley
from the Atomic Frontier.

141
00:07:20.130 --> 00:07:20.960
Keep looking up.

142
00:07:22.740 --> 00:07:24.120
- Check out Atomic Frontier!

143
00:07:24.120 --> 00:07:27.460
You could start with James's video
about a country-sized telescope

144
00:07:27.460 --> 00:07:29.900
or about the world's deepest scuba dive.
Next week,

145
00:07:29.900 --> 00:07:33.890
me. But also, not me.
It'll make sense, I promise.