Climate Exchange

[

] 135

isaster

isk

eduction

a higher level of skill than the climatological value of T7d for the

period up to two weeks ahead. It was also found that the prediction

skill for the probability of T7d values, at 20° C or below and 27° C or

above, was reasonable. JMA verified this in relation to past extreme

events, including the very high temperatures seen in 1994 and the

very low temperatures seen in 2003, and found realistic temperature

distributions and variations of the prediction comparable to actually

observed values.

Experimental provision of climate information

During summer 2011, JMA and NARO/TARC experimentally

provided the new climate information through a website to regis-

tered users (see map image on next page). In fact, NARO/TARC,

in collaboration with the Faculty of Software and Information

Science at Iwate Prefectural University (IPU) had operated a

website providing weather and related tailored information

aimed at reducing rice crop damage even before the pilot project

was implemented. Under the project, NARO/TARC further

customized information to users and improved the web system

to provide new climate information covering the period up to

two weeks ahead. The data included mean T7d predictions and

the predicted probability of T7d values at 20° C or below and

27° C or above for each 1 km grid, covering the period from one

to two weeks ahead. On the website, users could also view time-

sequence information for specific registered points, including

predicted T7d values and their variability (standard deviation)

and the predicted probability of T7d values at the levels described

above. The site also included maps showing the climatological

occurrence of the high and low temperatures and

related interpretation. When the predicted probability

of T7d reaching the alarming temperature was high

at a registered point, an alert was automatically sent

to users’ e-mail terminals and mobile phones. For

example, if the possibility of temperatures at or below

20° C was high, users received an alert so that they

could increase water levels in rice fields to protect

against low-temperature-related damage.

After the experiment of July 2011, NARO/TARC and

the IPU conducted an e-mail questionnaire survey. The

questionnaire was distributed to the 154 users who partic-

ipated in the experiment, and 89 replies were received.

Unfortunately (as far as the experiment is concerned),

no remarkably high or low temperatures occurred during

the trial period, meaning that few alert messages were

sent. Nonetheless, a number of respondents clearly indi-

cated the usefulness of the information provided during

the experimental period. Others underlined the need

for climate information covering the period up to two

weeks ahead. The results of the survey indicated that the

experimental provision was fruitful, in that it highlighted

agricultural users’ need for longer-range forecasts in the

Tohoku region and clarified that the provision of tailored

climate information offers potential benefits to farmers.

These findings were an important outcome of the pilot

project. However, as no extremely cold conditions were

seen during the experimental period in 2011, it is prema-

Seven-day mean temperature (T7d) prediction at a 1 km resolution in the Tohoku region of Japan

Source: JMA

(a) Climatological normal of T7d at

a 1 km resolution

(b) Mean value of predicted regional-scale

T7d anomalies

resolution (= (a) + (b))