1 INTRODUCTION

The usage of observational data to estimate physical conditions in astronomical objects has - as far as the calibration is concerned - two basic requirements. The data need to be calibrated, and the uncertainty in this calibration must be known quantitatively to the principal investigator (PI) of a science project. In radio astronomy, many different factors enter into the calibration. There are instrumental factors, like the absolute temperature scale, efficiencies, sideband ratios and such, as well as environmental factors, like atmospheric transmission and stability. In the sub-mm atmospheric window, which has just been opened during the last two decades for astronomical observations, latest receiver technologies are used, and the atmospheric absorption (and emission) plays a significant role in the data taking process. Thus for sub-mm observations these factors become even more important than for classical radio astronomy at longer wavelengths.

The Atacama Pathfinder EXperiment (APEX)[1, 2] is a modified ALMA prototype antenna, with 12 m diameter, used for frequencies between 200 GHz and 1.4 THz. APEX is currently equipped with two facility continuum receivers: The Large APEX Bolometer Camera (LABOCA)[3, 4] is a 295 channel bolometer array, for observations in the $\lambda\,850\,\mu{\rm m}$ atmospheric window. The Submillimeter APEX Bolometer Camera (SABOCA)[5] is a 37 channel TES bolometer array, for the $\lambda\,350\,\mu{\rm m}$ atmospheric window. For spectral line observations, the Swedish Heterodyne Facility Instrument (SHFI)[6] is available, with four single-pixel heterodyne receivers with central frequencies of 230, 345, 460, and 1300GHz.

The APEX telescope is located at an altitude of 5100 m in the Chilean Altiplano, close to the desert village San Pedro de Atacama. Its location aims to minimize the environmental effects on the calibration, as well as the use of the latest technology aims to minimize the instrumental effects. Still these effects exist, and they need to be taken into account during observations and data reduction to obtain properly calibrated science data.

At APEX we take these effects into account by the implementation of a calibration plan. The ultimate goal of this plan is to optimize the absolute intensity calibration of the obtained science data, and to quantify the remaining calibration uncertainty. The efforts to ensure a good calibration and to quantify it can be divided into two phases. The first phase comprises the observations (and their preparation), where real-time measures are taken in order to obtain the best possible calibrated data. The second phase comprises the data reduction, when the data are going through various calibration steps. This second phase depends strongly on the type of observation (continuum or spectral line), since different calibration schemes are applied and therefore different calibration measurements and procedures are required. Thus this paper is organized in the following way. In section 2 we will present the calibration efforts performed before or during the data taking process. Section 3 will describe the post-observation calibration efforts for continuum data, while section 4 will give an overview over the spectral line calibration. Both sections do not only descibe the calibration, but also address the problem of the calibration uncertainty of the final data. Finally, in section 5 we will summarize the achievements obtained with the calibration plan implementation at APEX.