system 
bers of 
issues 
> This 
recent 
diation 
press). 
1er and 
satellite 
or total 
ata set. 
lapping 
stantial 
o and 8 
R fora 
eraphic 
ocation 
CP-PL, 
nificant 
driving 
etations 
erences 
data set 
id other 
fit from 
ntative, 
remote 
lata set 
Development of Global Evapotranspiration Dataset and 
Land Cover Dataset of Asia 
Ryutaro Tateishi 
Center for Environmental Remote Sensing (CEReS) 
Chiba University 
Japan 
Global datasets of key environmental variables can be developed by the methods of (1)remote 
sensing, (2)ground measurement, (3)socioeconomical survey, and (4)the derivation from existing 
global datasets. Global evapotranspiration dataset was developed by the method (4), and Land cover 
dataset of Asia is being developed by the method (1) by the author. 
Global Evapotranspiration Dataset 
Monthly global land 30-minute grid datasets for potential evapotranspiration, evapotranspiration, and 
water surplus/deficit were developed(Ahn and Tateishi 1994a, 1994b). The Priestley-Tailor method 
was used to estimate potential evapotranspiration. The used data for the development of potential 
evapotranspiration are temperature by Legates and Willmott, albedo by Matthews, cloudiness by 
Leemans and Cramer, ETOPOS, and precipitation by Legates and Willmott. All these data are 
obtained from Global Ecosystems Data Base by NOAA-EPA. Monthly global land 30-minute 
evapotranspiration dataset was developed based on a simple water balance model using monthly 
precipitation, potential evapotranspiration and soil moisture data. Soil moisture data were derived by 
Donker's method using Water Holding Capacity data from UNEP/GRID-Geneva. The developed 
datasets will be available from GRID-Tsukuba(Fax: +81-298-58-2645), NGDC (Fax: +1-303-497- 
6513), or Rutgers University(Fax: 1-908-932-8746). 
The most difficult part of the development of evapotranspiration is validation. There is no truth data 
for 30-minute evapotranspiration. The author had to compare the developed data with estimated 
evapotranspiration data at a point location or 5-minute grid area by other researchers. The 
establishment of a standard methodology of accuracy/quality evaluation and a consensus of a standard 
description of accuracy/quality attached to developed datasets are important aspects for the 
development of better global databases. 
Land Cover Dataset of Asia 
Another topic of the presentation is a project to develop land cover dataset of Asia. The working 
group on "1-km Land cover Database of Asia” was established in the Asian Association on Remote 
Sensing (AARS) in October 1993 (Tateishi 1994). The working group consists of 49 members from 
28 Asian and Oceanian countries. Land cover dataset of Asia with 8-km resolution is planned to be 
produced early 1996 using NOAA/NASA Pathfinder AVHRR Land Data Set. After that, the 
resolution will be improved to 4km and 1km year by year. 
The working group's activities in 1995 consists of (1) modification of land cover legend, (2) ground 
truth collection, and (3) processing of 8-km AVHRR data. Land cover legend was established in 
1994 in the working group (Tateishi 1994). Modification of this legend is under going. Existing 
land cover/use maps with a scale of 1:1,000,000 (or 1:500,000 or 1:2,000,000) of each member 
countries will be collected and each member will collect ground truth based on the determined land