The study describes the variation in physical and chemical data from a trial with carrots grown outdoors and in plastic greenhouses during two years at three geographical locations in Norway, 59°40’N, 63°28’N and 69°39’N, and analysed by principal component analysis (PCA), partial least square and multiple linear regression (MLR). The most important principal component (PCI) to describe the total variation was the root weight, root diameters, and the carotene content. PCI described 53–65% of the total variation. Further 13–17% of the variation could be explained by a second component which included the air temperature a fortnight prior to harvest, and the sugar content of the roots. PLS and MLR revealed relationships between chemical and physical variables. Physical variables were able to predict 55% of the variation in chemical variables by the first two PLS components. The biological development of the carrot roots was best described by the root weight and a computed variable containing root dry weight and cylindricity of the carrot roots.