Matching LOFAR sources across radio bands

Aims. The recent preliminary release of the LOFAR LBA Sky Survey (LoLSS) is the first wide-area, ultra-low frequency observation published from the Low Frequency Array (LOFAR). Our aim is to combine this data set with other surveys at higher frequencies to study the spectral properties of a large sample of radio sources. Methods. We present a new cross-matching algorithm that takes into account the sizes of the radio sources. We applied it to the LoLSS-PR, LoTSS-DR1, LoTSS-DR2 (all LOFAR), TGSS-ADR1 (GMRT), WENSS (WSRT), and NVSS (VLA) catalogues. We then studied the number of matched counterparts for LoLSS radio sources and their spectral properties. Results. We found counterparts for 22 607 (89.5%) LoLSS sources. The remaining 2640 sources (10.5%) were identified either as an artefact in the LoLSS survey (3.6%) or flagged due to their closeness to bright sources (6.9%). We found an average spectral index of α = −0.77 ± 0.18 between LoLSS and NVSS. Between LoLSS and LoTSS-DR2, we found α = −0.71 ± 0.31. The average spectral index is independent of the flux density above S54 = 181 mJy. A comparison of the spectral slopes from LoLSS-LoTSS-DR2 with LoTSS-DR2–NVSS indicates that the probed population of radio sources shows evidence of a negative spectral curvature.