Long-term change and spatial anomaly of

1
Long-term change and spatial anomaly of warm
season afternoon precipitation in Tokyo
Fumiaki Fujibe1, Hiroki Togawa2 and Masaaki
Sakata2 1Meteorological Research Institute,
Japan 2Tokyo District Meteorological Observatory,
Japan
AMS 89th Conference
12 Jan. 2009
2

Urban flood disaster (Kobe, 28 July 2008)
Five people (including three children) were
killed by a sudden surge of water in a river.
Radar echoes
1400 JST
1416 JST
1438 JST
Kobe
100 km
1442 JST
1444 JST
mm/hour
http//www.kobe-np.jp/rentoku/jiken/200807togagawa
/ 0001372826.shtml
http//www.osaka-jma.go.jp/saigai/
pdf/h20/sokuji/Rain0728.pdf
3

Urban flood disaster (Tokyo, 5 August 2008)
Five workers were killed by a sudden surge of
water in a sewage tunnel.
Radar echoes
1150 JST
25 km
Tokyo Bay
Toshima ward
Toshima Ward
134 mm
mm/hour
http//www.tokyo-jma.go.jp/ sub_index/bosai/disast
er/ 20080805/20080805.html
Daily precipitation
112 mm
91 mm
15 km
20 40 60 80 100 120 mm
4

Questions on urban precipitation
(1) Is heavy rainfall increasing at cities? (2)
Does the urban heat island increase rainfall?
The heat island of Tokyo is indeed very
conspicuous. However, it is not easy to give
convincing evidence for precipitation change.
The left picture was taken from one of these
windows.
JMA Building
5

Features of the Tokyo heat island
gt1000m m.s.l.
Population density 1000 3000 10000 km-2
Tokyo 3.1?/century
Tokyo (Otemachi)
Departure from 1906-1925 average
30.7 million people in a 50km radius
Global mean 0.8?/century
Hachijo Is. 0.6?/century
Hachijo Is.
Cited from Yamashita (1996, Atmos. Environ.)
6

Urban effects on daytime cloud amount
Daytime cumulus amount (NOAA) for April-October
(1991-2001)
NDVI pattern (LANDSAT)
urban area along railways
Low
High
Low
High
More cloud amount along low NDVI (urbanized)
area that develops along railways


Published by Inoue and Kimura (2004, GRL)
7

Long-term heavy precipitation change at Tokyo
Annual maximum hourly precipitation
mm
Multidecadal variations rather than linear
increase

Cited from Kanae et al. (2004, JMSJ)
8

Questions on urban precipitation
(1) Is heavy rainfall increasing at cities? (2)
Does the urban heat island increase rainfall?
The heat island of Tokyo is indeed very
conspicuous. However, it is not easy to give
convincing evidence for precipitation change.

• Heavy rainfall in Japan is mainly caused by
  large-scale disturbances (such as typhoons and
  Baiu fronts) that are not likely to be affected
  by the urban heat island.
• Urban influence is most likely to be enhanced for
  short-term, localized showers on sunny days in
  the warm season.
• Analysis should be focused on short-term
  precipitation in the warm season.

9
Outline of the study
Data

• Digitization of hourly data
• Precipitation at Tokyo since 1890.
• Full quality check through one-by-one comparison
  with the original record.

Analysis

• Long-term precipitation change
• For no preceding precipitation (NPP) cases,
  defined as not preceded by 1mm precipitation for
  six hours.

Supplementary analysis

• Spatial anomaly of precipitation in Tokyo
• Based on hourly data for 1978-2007.
• Check the consistency between long-term trends
  and spatial anomaly.

10

Precipitation trend for NPP cases at Tokyo
1700 to 2300 local time, June to August
Eleven-year running mean
No preceding precip. (NPP)
All cases
Linear trend 729/century
Linear trend 4847/century
11

Precipitation trend for NPP cases at Tokyo
Significant at 5
Significant at 1
No preceding precip. (NPP)
All cases
12

Precipitation trend for NPP cases (for frequency)
Significant at 5
Significant at 1
1mm/(6h) frequency
3mm/(6h) frequency
5mm/(6h) frequency
10mm/(6h) frequency
20mm/(6h) frequency
13

Spatial precipitation distribution for NPP cases

Data Hourly precipitation on the AMeDAS
network Period1978-2007 (30 years) Definition of
NPP cases No precipitation (1mm) for the
preceding six hours at 90 among 62 stations in
the area (shown in ?)
Spatial interpolation from surrounding
stations SPi-(Paxibyicxi2dyi2exiyi)2?
min., i Definition of spatial anomaly
Q ltP0gt/lt Pgt -1
Nerima
Otemachi
Setagaya
1700 to 2300 local time, June to August
14

Spatial anomaly for NPP cases (Tokyo, Otemachi)
Significant at 5
Significant at 1
No preceding precip. (NPP)
All cases
15

Spatial anomaly for NPP cases (other stations)
Significant at 5
Significant at 1
Nerima
Setagaya
16

Summary
1. Hourly precipitation data at Tokyo from 1890
were digitized and quality-checked. 2. For no
preceding precipitation (NPP) cases,
precipitation amount at Tokyo shows an increasing
trend of 30/century or more in the evening of
the warm season. 3. Analysis for spatial
precipitation distribution for NPP cases also
shows a positive precipitation anomaly in Tokyo
in the afternoon of the warm season, in
consistency with the feature of the long-term
trend.
17


18

Precipitation trend (NPP defined for different
duration)
Significant at 5
Significant at 1
t 3 hours
t 9 hours
t 1 hour