Laboratory for
Climatology and Remote Sensing

Abstract:

Quality control is a particularly demanding problem for micrometeorological studies in complex environments. With the transition to electronic sensing and storage of climate data in high temporal resolution, traditional approaches of homogenization are insufficient for addressing the small-scale variability and spatial heterogeneity of the data. This problem can be successfully addressed by introducing a new class of control procedures based on the physical and climatological relations between different climate variables. The new approach utilizes knowledge about the interdependency of air temperature, precipitation, radiation, relative air humidity, cloud cover and visibility to develop empirical functions for determining the probability margins for the co-occurrence of specific conditions in tropical mountains and deserts. It can also be applied to other geographic settings by adjusting the parameters derived from the data itself. All procedures are integrated into a processing chain with feedback loops and combined with conventional logical and statistical checks, which enables it to detect small errors that normally pass unnoticed. The algorithms are also adapted to incorporate the short time steps of the original data to retain the potential for detailed process analyses.

Abstract:

One crucial aspect within most joint projects in ecosystem research is the way of data storage and exchange. The Data Warehouse of the DFG-Research Unit 816 (FOR816dw) provides an integrative platform to enable the retrieval, transfer, storage, and analysis of scientific data of various ecological disciplines, focused on biodiversity research. To structure the data and to make it usable by other scientists a robust metadata concept had to be implemented. The Ecological Metadata Language (EML: www.knb.ecoinformatics.org/software/eml) provides a sophisticated XML-based scheme to describe scientific ecological data sets. The relational database structure of the FOR816dw is based on the metadata logic of EML and contains at least all mandatory fields to be compliant to this specification. The metadata are transmitted by the data owner via a browser based wizard during the data upload. They can be queried for temporal, spatial, and thematically features to find data sets in the data pool. Furthermore the metadata bears information on the data set creator, the intellectual rights, and the data collection/processing workflow of each contained attribute of a data set. The adaption of the concept to describe each attribute (variable) of a data set establishes the granulated storage and analysis of the tabular data set values. All information are safely stored in the relational database and can be filtered, selected and reused. Beside the storage and maintenance of data sets and metadata the FOR816dw provides the management of administrative tasks of the research group. The user management, the project structure, a news system, a mailing-lists generator, the travel reimbursement, and the station booking system is based on the same personnel and project specific data used for the metadata description of the scientific data sets. The project`s webpage (www.tropicalmountainforest.org) is the single point of contact for the project members and provides a user-friendly access to all information of the DFG-Research Unit 816 ?Biodiversity and Sustainable Management of a Megadiverse Mountain Ecosystem in South Ecuador?. After six years of development and running, an open source release will be available at the termination of the FOR816 in March 2013. With this in view we want to review some technical and conceptual solutions and discuss the acceptance by the users. Is the representation of EML in a relational database possible? Is the combination of project administration and data management in one platform useful? Up to what level of detail are the users willing to submit and describe their data?