Management practices, geographical gradients and climatic factors are factors explaining weed species composition and richness in cereal fields from Northern and Central Europe. In the Mediterranean area, the precise factors responsible for weed distribution are less known due to the lack of data and surveys. The existence of weed survey data of year 1976 in the Zaragoza province of the Aragon region, Spain, offered us the opportunity to compare present weed species with weed species growing 30 years ago. No detailed comparison of changes in weed species composition in cereal fields in that period of time has been conducted in the Mediterranean area. Here a survey was conducted in the Aragon region from 2005 to 2007. Weeds were surveyed in 138 winter cereal fields in ten survey areas where winter cereals are the main crops, using the same methodology applied 30 years ago. In the Zaragoza province, 36 fields were chosen in the same municipalities than in the previous survey. Several management, geographic and climatic variables of each field were recorded and related to weed species with multivariate analysis. Diversity index were calculated and related to survey area and altitude. Our results show that out of the 175 species only 26 species were found in more than 10% of the surveyed fields. The main species were Papaver rhoeas, Lolium rigidum, Avena sterilis and Convolvulus arvensis found in more than half of the surveyed fields. L. rigidum was related to dryland, while the other species were found overall. Furthermore, we found that management, geographical and climatic factors were significantly related to weed species distribution. In particular altitude, survey areas, irrigation and herbicide use in post-emergence were the most driving factors explaining weed species distribution. Species richness was higher in survey areas with extensive management practices and increased with altitude excepting a very productive area with intensive management practices at high altitude where richness was as low as in the irrigated lowlands. The main differences found between the 1976 and the 2005-2007 surveys were (1) the striking increase of grass weeds, (2) the high decrease of mean weed species number found in each field declining from 9 to 3 and (3) the frequency decrease of many weed species probably caused by agriculture intensification in that period of time. The growing importance of other weed species is probably related to their adaptation to minimum tillage, which is a widespread technique nowadays.
