al 
SATELLITE- AND GROUND-BASED MULTI-VIEW PHOTOGRAMMETRIC 
DETERMINATION OF 3D CLOUD GEOMETRY 
G. Seiz', D. Poli, A. Gruen, E. P. Baltsavias, A. Roditakis 
Institute of Geodesy and Photogrammetry, Swiss Federal Institute of Technology (ETH) Zurich 
ETH Hoenggerberg, CH-8093 Zurich, Switzerland 
(gseiz, daniela, agruen, manos, roditak } @geod.baug.ethz.ch 
Commission VII, WG VII/6 
KEY WORDS: Atmosphere, Global Change, Sensor Orientation, Matching, Camera, Visualization. 
ABSTRACT: 
The quantitative 3D description of clouds is important for refined methods in nowcasting and the modeling of weather and climate. 
The EU project Cloudmap aimed at developing new methodologies for cloud product derivation (heights, type, optical thickness and 
effective droplet size). The follow-up project Cloudmap2 aimed at producing and exploiting value-added remote sensing data 
products on macroscopic (e.g. cloud-top height) and microscopic (e.g. cloud droplet radius) properties and water vapour 
distributions in order to characterize sub-grid scale processes within Numerical Weather Prediction Models (NWP) through 
validation and data assimilation. 
Earth Observation (EO) image data, provided by ESA, EUMETSAT and NASA are used to derive geophysical value-added data 
products over Europe and the North Atlantic region, whenever possible in near real-time. Ground-based active (cloud radar, 
ceilometer) and passive (stereo imager system, IR camera) remote-sensing instruments are used to validate the EO-derived products 
as well as to merge them with the satellite-based results for a full 3D representation of the clouds. The role of our group in 
Cloudmap2 was to estimate cloud-top height (CTH) and wind (CTW) from stereo images from satellites and cloud-bottom height 
(CBH) and wind (CBW) from stereo images acquired by our newly developed ground-based stereo imager system. The cloud-top 
and -bottom results were then combined into a 3D model and visu 
alized. This paper describes the results obtained in CTH and CTW 
estimation from ATSR2, AATSR, MISR and Meteosat-6/-7, including validation, the CBH and CBW results from the ground-based 
stereo imager system and a case study where the satellite- and ground-based 3D cloud boundary results are combined. 
1. INTRODUCTION 
The interest to monitor cloud properties from space and ground- 
based observations is based on the large influence that clouds 
have on the Earth and Atmosphere energy balance. The EU-FP4 
project Cloudmap aimed at estimating new cloud-top products 
(heights, type, optical thickness, effective droplet size), 
especially for cirrus and contrail clouds from existing and new 
sensors, using three different techniques (brightness temperature 
with CO2 slicing method, stereoscopy and Oxygen A-band). 
These cloud-top products have been validated using airborne 
sensor underflights, multi-resolution observations from space 
sensors and ground-based remote sensing instruments. 
Cloudmap ended in January 2001 and was then continued by 
the EU-FPS project Cloudmap2 (www- 
research.ge.ucl.ac.uk/cloudmap2/) until June 2004. 
Cloudmap2 aimed at producing and exploiting value-added 
remote sensing data products on macroscopic (e.g. cloud-top 
height) and microscopic (e.g. cloud droplet radius) properties 
and water vapour distributions in order to characterize sub-grid 
scale processes within Numerical Weather Prediction Models 
(NWP) through validation and data assimilation. Earth 
Observation (EO) data, provided by ESA, EUMETSAT and 
NASA are used to derive geophysical value-added data 
products over Europe and the North Atlantic region, whenever 
possible in near real-time. Ground-based active (cloud radar, 
ceilometer) and passive (stereo imager system, IR camera) 
  
: Corresponding author. 
remote sensing instruments are used to validate the EO-derived 
products as well as to be merged with the satellite-based results 
for a whole .3D representation of the clouds. Numerical 
simulation experiments based on state-of-the-art radiative 
transfer methods are used to quantify the effect of broken clouds 
on the Earth's radiation budget and lead to a better 
representation of clouds within NWP models. 
The role of our group in Cloudmap and Cloudmap2 was to 
estimate cloud-top height (CTH) and wind (CTW) from stereo 
images from satellites and cloud-base height (CBH) and wind 
(CBW) from stereo images acquired by our own ground-based 
stereo imager system, with stereo-photogrammetric techniques. 
As second step, the cloud-top and -base results were then 
combined and visualized in 3D. 
This paper describes the CTH and CTW retrieval from multi- 
view satellite sensors (ATSR2/AATSR, MISR) and 
geostationary satellites (Meteosat-6/-7) and the 
comparison/validation of these cloud products with other 
satellite-based products as well as measurements from ground- 
based instruments (multi-camera system, cloud radar). Finally, a 
case study is presented where the satellite- and ground-based 
3D cloud boundary results were combined. 
1155