Context: The dusty debris disk around the $sim$20 Myr old main-sequence A-star {beta} Pictoris is known to contain gas. Evidence points towards a secondary origin of the gas as opposed to being a direct remnant form the initial protoplanetary disk, a
lthough the dominant gas production mechanism is so far not identified. The origin of the observed overabundance of C and O compared to solar abundances of metallic elements, e.g. Na and Fe, is also unclear. Aims: Our goal is to constrain the spatial distribution of C in the disk, and thereby the gas origin and its abundance pattern. Methods: We used the HIFI instrument onboard Herschel to observe and spectrally resolve CII emission at 158 $mu$m from the {beta} Pic debris disk. Assuming Keplerian rotation, we use the spectrally resolved line profile to constrain the spatial distribution of the gas. Results: We show that most of the gas is located around $sim$100 AU or beyond. We estimate a total C gas mass of $1.3times10^{-2}$ M$_oplus$. The data suggest that more gas is located on the southwest side of the disk than on the northeast side. The data are consistent with the hypothesis of a well-mixed gas (constant C/Fe ratio throughout the disk). Assuming instead a spatial profile expected from a simplified accretion disk model, we found it to give a significantly worse fit to the observations. Conclusions: Since the bulk of the gas is found outside 30 AU, we argue that the cometary objects known as falling evaporating bodies are unlikely to be the dominant source of gas; production from grain-grain collisions or photodesorption seems more likely. The incompatibility of the observations with a simplified accretion disk model could favour a preferential depletion explanation for the overabundance of C and O. More stringent constraints on the spatial distribution will be available from ALMA observations of CI at 609 $mu$m.
