Cloud Mapping from the Ground: Use of Photogrammetric Methods

from the existing observation networks, except at a very few Abstract places around the world which are equipped with high-cost inWithin the European Union (EU) project Cloudmap, a groundstruments (see Table 1). At most climate stations of the national based sky imager system consisting of two commercial digital networks, cloud macroscopic properties—mainly cloud cover, CCD cameras with wide-angle lenses has been established which cloud depth, and cloud-base height—are still visually obcan theoretically be used to derive various macroscopic cloud served. Mainly at airports, ceilometers are in operational use to parameters (cloud-base height, cloud-base wind, cloud measure the cloud-base height automatically and continuously amount). In this paper, we present the method to calculate a in addition to the visual observations. It is well recognized today digital surface model (DSM) of the cloud base. It includes both that the infrequent, spatially not equally distributed, subjective, the precise determination of the interior and exterior orientaand too sparse point observations of clouds do not meet the retion of the cameras as well as the automatic derivation of the quirements of numerical weather prediction (NWP) and global cloud-base heights using modern photogrammetric algorithms. climate models (GCM). The data from our ground-based stereo The presented measurements were taken during the imager system will in this context nicely complement existing Mesoscale Alpine Programme (MAP) in Switzerland in October instruments and observations with very accurate, spatial, and 1999. The results from our own matching software and from frequent 3D cloud-base heights which should be suitable for the commercial photogrammetric systems were validated with meteorological models. semi-automatically measured points and compared with visual Within the EU project Cloudmap2 (Cloudmap2, 2001), we observations, lidar, and radiosonde data from the MAP are currently working on gaining knowledge about the requireComposite Observing Network and satellite-based cloud-top ments of temporal frequency, spatial coverage, and accuracy for heights from ERS2-ATSR2. The potential of the system to provide cloud products to be used as assimilation data into NWP or spevery accurate areal cloud-base height data was shown. This is cific cloud models. These new requirements are certainly important for the objectives of the EU project Cloudmap2, where higher than the operational user requirements defined in the it is planned to assimilate various cloud parameters, including World Meteorological Organization (WMO) Guide to Meteorocloud-base height, into cloud and high-resolution numerical logical Instrumentation and Methods of Observation, which weather prediction (NWP) models. state for CBH a vertical accuracy of 10 percent, a height range from 30 m to 30 km, but no specific numbers about time resoluIntroduction tion or area coverage (Hans Roozekrans, personal communicaClouds play a pivotal role in the interaction between the Earth’s tion). In particular, areal measurements or much denser netclimate and anthropogenic inputs, particularly from the inworks will possibly be required. crease in greenhouse gases. Therefore, accurate global measureGround-based digital cameras in stereo configuration can ments of the relative location, distribution, and character of help to fulfil some of the user requirements described above. clouds (which have a strong impact on both the total incoming Analog stereo images of clouds have been taken and analyzed radiation at the surface and the reflected radiation above the for more than a hundred years (Koppe, 1896). It was shown that cloud field) are necessary as described in the rationales of the cloud-base height and cloud-base motion can be determined European Union (EU) projects Cloudmap and Cloudmap2 from such systems and used as validation data for satellite(Cloudmap, 2001; Cloudmap2, 2001). For receiving global covbased cloud properties (Bradbury and Fujita, 1968). Nevertheerage, satellite-based methods have to be used, but they have to less, such an analog system is only useful for research purbe calibrated and validated with ground-based measurements; poses; any operational use of the analog images was not possifurthermore, the global numerical weather prediction (NWP) ble due to the enormous amount of time to analyze a single and climate models are refined by higher-resolution regional time step by first scanning the images and then finding the cormodels whose initial conditions have again to be updated by responding points by manual measurements. Digital systems ground-based measurements. In Cloudmap, three independent have the major advantage of reducing the processing time sigground-based validation instruments were operated within varnificantly, down to less than an hour, which gives them the poious measurement campaigns: cloud radars, airborne lidar profitential of deriving the cloud parameters in near real time. Praclometers, and stereo photogrammetric systems such as the one tical fieldwork with monoscopic digital whole-sky imager described in this paper. While the first two instruments are also systems (WSI) (Shields et al., 1999) was first performed at variused to derive microphysical cloud properties, our instrument ous Atmospheric Radiation Measurement (ARM) Program sites was tested for delivering very accurate areal data of the 3D cloud base, which are necessary for weather and climate models, as described later in this paragraph. Cloud-base height (CBH) is one important factor in determining the infrared radiative properties of clouds. However, cloud-base heights are not well known Photogrammetric Engineering & Remote Sensing