Climate Exchange

[

] 136

isaster

isk

eduction

ture to conclude the effectiveness of the new climate information

services and identify possible related problems. JMA and NARO/TARC

will continue to supply the new information on a trial basis in summer

2012 towards its operational provision in the near future.

Success factors

The pilot project was successful both for JMA and for NARO/TARC.

In particular, JMA gained significant expertise on developing better

climate information services. Some keys to the initiative’s success

and lessons learnt from it are summarized below:

• It is critical for climate information users and providers to

engage in close and productive dialogue and to share knowledge

due to the nature and difficulty of climate prediction and

information. JMA and NARO/TARC frequently shared relevant

information both by e-mail and at face-to-face meetings, which

supported the success of the project.

• Intermediaries linking providers and end users of climate

information are very important in promoting its use. In the pilot

project, NARO/TARC and its researchers acted as intermediaries

between JMA and end users in the agricultural sector.

• Effective use of existing systems, as well as support from partners,

contributes to the smooth and successful launch of new services.

In the project, experimental provision of the new information was

effectively carried out using existing systems operated by NARO/

TARC in collaboration with the IPU, which offered useful support

as it specializes in information technology. Making the most of its

experience to implement the pilot project, NARO/TARC took steps to

improve its agro-meteorology information based on user feedback.

Future plans and expectations

Based on the success of the pilot project with NARO/TARC, JMA held

discussions with NARO (the higher authority of NARO/TARC) with a

view to efficiently expanding such collaboration to initi-

ate similar projects in other regions. Currently, JMA is

involved in four pilot projects in conjunction with four

regional agricultural research centres under NARO. In

addition, recognizing the effectiveness of tailored climate

information in the agricultural sector, JMA has begun

discussions with Japan’s Ministry of Agriculture, Forestry

and Fisheries (which oversees and connects all stake-

holders in the agricultural sector, including those within

local governments) to develop better partnerships with

the sector and promote further development and effective

use of climate information services beneficial to various

agricultural activities.

JMA is further seeking opportunities to collaborate with

operators in other fields, including the energy sector, in

order to provide greater levels of usefulness and tailoring

in its climate information services. Such collaboration is

expected to promote efficient climate information services

that are beneficial in a variety of socioeconomic activities.

A map for the experimental provision of temperature

prediction to specific users (in Japanese)

Source: JMA

Temperature prediction at a 1 km resolution

T7d prediction at a 1 km resolution was made based on

a combination of two data sets: NARO/TARC provided

climatological normals of T7d at a 1 km resolution

covering the 30-year period from 1981 to 2010, while

JMA provided PDF prediction data for regional-scale

T7d anomalies for the Sea of Japan side and the Pacific

Ocean side of the Tohoku region.

The climatological normals of T7d in more than

70,000 1 km grids in the Tohoku region were calculated

from historical records of daily mean temperatures

(T1d), which were statistically estimated using surface

observation data of T1d and geographical information at

a 1 km resolution. These observation data were obtained

from a network of automated weather stations operated

by JMA, known as the Automated Meteorological Data

Acquisition System (AMeDAS).

JMA provided predicted PDFs of regional-scale T7d

anomalies that are made twice a week for EWIEW. The

PDFs were calculated using both numerical prediction

data from JMA’s operational one-month Ensemble

Prediction System (EPS) and regression coefficients

estimated from hindcast experiments (re-forecasts)

for 1981-2010, which were carried out with the same

system as that in operational use. To calculate the

predicted PDFs of T7d at a 1 km resolution with respect

to actual temperatures (rather than in three operational

categories), two steps must be completed: statistical

downscaling for predicted PDFs of T7d anomalies from a

regional scale to a 1 km resolution, and addition of the

PDFs to the climatological normal.

When the predicted PDFs were changed from a regional

scale to a 1 km resolution, the mean values were

assumed to be uniform and equal to that in each region.

However, PDF variances are thought to differ for each 1

km grid even in the same region. Accordingly, differences

in variances caused by the different spatial scale were

estimated using historical records of temperature at

a 1 km resolution provided by NARO/TARC. Based on

this estimation, PDFs of the T7d anomaly at a 1 km

resolution were calculated from the regional-scale

values. After this process, the PDFs were added to the

climatological normals for each 1 km grid so that users

could use the PDFs in view of actual temperatures rather

than anomalies. This facilitated user understanding and

supported use in decision-making.

Users can select mean values of the seven-day mean temperature (T7d) and the

predicted probability of T7d values at or below (above) 20 (27)° C at a 1 km resolution

for the period from one to two weeks ahead of 27 July, 2011, on the website