1
00:00:00,130 --> 00:00:04,200
VO: In order improve hurricane forecasts, scientists model storms of the past.

2
00:00:04,220 --> 00:00:08,310
Braun: We model old storms because it gives a way to

3
00:00:08,330 --> 00:00:12,470
compare the simulations to observations and try to assess the

4
00:00:12,490 --> 00:00:16,630
physical processes within storms and also how well those

5
00:00:16,650 --> 00:00:20,760
processes are represented in the forecast models. Because that's key to

6
00:00:20,780 --> 00:00:24,940
improving forecasts in the future.
VO: With improvements in computing power, 

7
00:00:24,960 --> 00:00:28,970
researchers have been running ensemble forecasts.

8
00:00:28,990 --> 00:00:33,050
Braun: Instead of a single forecast, we run a whole series of forecasts where you make slight

9
00:00:33,070 --> 00:00:37,160
variations to the initial information that goes into the models and then see how much

10
00:00:37,180 --> 00:00:41,220
spread you get within those forecasts.
Running ensembles gives us a visual description

11
00:00:41,240 --> 00:00:45,350
of the level of uncertainty associated with forecasting tropical storms.

12
00:00:45,370 --> 00:00:49,460
Braun: And the average over all those ensembles usually gives you a better forecast

13
00:00:49,480 --> 00:00:53,580
than if you ran just a single forecast model.
VO: One way to test forecast models is to

14
00:00:53,600 --> 00:00:57,670
reanalyze past storms.
Reale: It's very important to study the

15
00:00:57,690 --> 00:01:01,840
worst possible storms in history because those are the ones that really put the

16
00:01:01,860 --> 00:01:05,990
system to a challenge.
VO: NASA and NOAA cooperate in satellite

17
00:01:06,010 --> 00:01:10,190
systems and sharing data, as well as experiments and modeling research, all of

18
00:01:10,210 --> 00:01:14,380
which enable NOAA and other agencies to provide better forecasts of tropical cylones.

19
00:01:14,400 --> 00:01:18,460
Reale: When NOAA and National Hurricane Center issues a forecast, the better the

20
00:01:18,480 --> 00:01:22,550
forecast is historically, the more likely it is that people would trust the

21
00:01:22,570 --> 00:01:26,640
forecast, and decision makers will make the right decisions and

22
00:01:26,660 --> 00:01:30,720
they will tell what place has to be evacuated and what place can stay.

23
00:01:30,740 --> 00:01:34,840
VO: To get a clearer look at the processes inside the hurricane, NASA's Global

24
00:01:34,860 --> 00:01:38,970
Precipitation Measurement Microwave Imager, or GMI, gives us an

25
00:01:38,990 --> 00:01:43,050
X-ray beneath the clouds.
Braun: If you have a storm that maybe isn't well defined, doesn't have

26
00:01:43,070 --> 00:01:47,210
a visible eye, with the GMI and similar instruments

27
00:01:47,230 --> 00:01:51,360
you're able to see the rainfall structure underneath those clouds

28
00:01:51,380 --> 00:01:55,440
So you can see the ring of heavy precipitation in the eye wall, as well as

29
00:01:55,460 --> 00:01:59,510
several of the rain bands. And that really helps to tell us a lot about how the storm is evolving,

30
00:01:59,530 --> 00:02:03,590
and how that might relate to the intensity of the storm.

31
00:02:03,610 --> 00:02:07,670
VO: Representing those small-scale eye processes in global models has been an ongoing

32
00:02:07,690 --> 00:02:11,780
area of research.
Reale: By increasing resolution and increasing, of course, the quality of

33
00:02:11,800 --> 00:02:15,910
the physical processes that are represented inside the model,

34
00:02:15,930 --> 00:02:19,990
global models have been getting better and better. They make hurricanes

35
00:02:20,010 --> 00:02:24,090
smaller, more compact, closer to the real size,

36
00:02:24,110 --> 00:02:28,280
and the eye becomes smaller and smaller and they become more intense.

37
00:02:28,300 --> 00:02:30,813