1
00:00:02,920 --> 00:00:07,080
This project is based on some work
by a PhD student of mine, Pete Tuckett,

2
00:00:07,080 --> 00:00:09,480
and the rationale was that

3
00:00:09,480 --> 00:00:12,880
we've got two problems with
mapping surface water on Antarctica,

4
00:00:12,880 --> 00:00:16,280
one being the scale of Antarctica
and another being clouds.

5
00:00:16,280 --> 00:00:20,320
So, we wanted to develop a method
that could solve both those problems.

6
00:00:20,320 --> 00:00:22,960
And so we used some open source software,

7
00:00:22,960 --> 00:00:26,000
Google Earth Engine,
which allows us to process

8
00:00:26,000 --> 00:00:29,440
huge amounts of satellite data
using Google's powerful servers.

9
00:00:29,440 --> 00:00:33,920
I think a key aim of the study was to produce a method
that could be used and developed by others.

10
00:00:33,920 --> 00:00:37,800
So, we developed and applied it
to one particular area of Antarctica

11
00:00:37,800 --> 00:00:41,280
with the overall aim of upscaling it
to the whole of Antarctica.

12
00:00:41,280 --> 00:00:44,280
We really want it to be an approach
that can be used by others

13
00:00:44,280 --> 00:00:45,240
and not just Antarctica.

14
00:00:45,240 --> 00:00:48,240
You could apply it to other ice masses
like Greenland, for instance.

15
00:00:48,240 --> 00:00:53,600
We wanted it to be in the FAIR principles.

16
00:00:53,600 --> 00:00:57,160
Google Earth Engine is freely available

17
00:00:57,160 --> 00:00:59,440
so the code was written in Google Earth Engine.

18
00:00:59,440 --> 00:01:03,160
We made sure that we've got clear metadata
and it was clearly commented

19
00:01:03,160 --> 00:01:04,600
for other people to use,

20
00:01:04,600 --> 00:01:06,960
take that data and use it for themselves.

21
00:01:06,960 --> 00:01:09,040
I don't think we suffered
too much in the way of barriers.

22
00:01:09,040 --> 00:01:13,000
I think actually, in physical geography
and in general glaciology

23
00:01:13,000 --> 00:01:17,000
there has been a big move
to making data open and accessible.

24
00:01:17,000 --> 00:01:20,560
We published in The Cryosphere
which is an open access journal.

25
00:01:20,560 --> 00:01:24,800
They have their own policy around
making sure data is accessible.

26
00:01:24,800 --> 00:01:27,280
I would say that actually, we were
almost encouraged to do so 

27
00:01:27,280 --> 00:01:29,400
as part of the publishing process.

28
00:01:29,400 --> 00:01:33,800
Really, that’s what we wanted: to be able to make it
so that others can make use of it.

29
00:01:33,800 --> 00:01:37,480
So in retrospect,
yeah, I think we could have done

30
00:01:37,480 --> 00:01:39,440
some of the processing
in other freely available software.

31
00:01:39,440 --> 00:01:43,360
So Google Earth Engine is where
we produce the code to map surface water,

32
00:01:43,360 --> 00:01:47,320
then we applied post-processing to try and clean it up

33
00:01:47,320 --> 00:01:49,800
and extract various metrics of what we wanted.

34
00:01:49,800 --> 00:01:51,440
 That was done in MATLAB.

35
00:01:51,440 --> 00:01:55,040
MATLAB does require a license,
so it's not accessible to everyone.

36
00:01:55,040 --> 00:01:59,040
So I guess looking back,
we could have also tried to do that

37
00:01:59,040 --> 00:02:02,040
in a language that could be used by all.

38
00:02:02,040 --> 00:02:06,800
I think FAIR principles are going to be
fundamental to my research going forward.

39
00:02:06,800 --> 00:02:10,040
The FAIR principles are key
to being able to produce

40
00:02:10,040 --> 00:02:14,680
or make our data and code available
to as wide a range of people as possible.

41
00:02:14,680 --> 00:02:19,560
We want to build on what other people have done
and tap into the big data that now exists.

42
00:02:19,560 --> 00:02:23,920
I think FAIR principles is something
that should be relatively straightforward

43
00:02:23,920 --> 00:02:27,560
to do now, or is becoming
increasingly accessible in itself

44
00:02:27,560 --> 00:02:30,600
just because of the push
from journals to do it.

45
00:02:30,600 --> 00:02:34,720
You get help when you're publishing
in terms of having publishable statements.

46
00:02:34,720 --> 00:02:39,720
There are also repositories that allow us
to store data and generate DOI’s associated with it,

47
00:02:39,720 --> 00:02:44,080
like the Sheffield repository 
that we've put our data in.

48
00:02:44,080 --> 00:02:47,920
My advice to other researchers
interested in making their data FAIR

49
00:02:47,920 --> 00:02:50,440
would firstly be that 
there's lots of support out there.

50
00:02:50,440 --> 00:02:54,680
I think the second point is that I think there's been

51
00:02:54,680 --> 00:02:58,920
a clear step towards seeing the 
importance of FAIR data

52
00:02:58,920 --> 00:03:02,040
to that collaborative approach
of generating knowledge.

53
00:03:02,040 --> 00:03:05,800
So, when I think of science,
I think of it as little building blocks,

54
00:03:05,800 --> 00:03:09,960
building towards making these little
advances which build towards other advances

55
00:03:09,960 --> 00:03:14,640
and being able to share your data
allows other people to use it

56
00:03:14,640 --> 00:03:18,160
and build on it and take your code
and apply it to different areas.

57
00:03:18,160 --> 00:03:21,640
So I think it's a fundamental part
of that progress in science

58
00:03:21,640 --> 00:03:24,520
and allowing us to make progress
in lots of different areas.