The 2002 drought across Colorado was equal, to if not more severe than, drought conditions over the previous 100 years. Drought conditions started appearing in parts of Colorado in the Fall of 1999 and this multiyear drought has persisted through much of 2004. As the primary user of water in the state, agriculture has been impacted significantly by this drought. The following study focuses on the role of climate information in decision making at the ranch and farm level with the goal of determining when and how climate information is most useful to these producers. The overarching framework was to elicit information on data and forecast quality, acceptability, context of use and accessibility, and an economic analysis was also carried out. Our research builds on a joint surveying project between Colorado State University and the NOAA Climate Diagnostics Center to collect information on 2002 drought impacts on agriculture in Colorado. At the end of the 2002 growing season, agricultural producers across Colorado were surveyed to determine drought impacts, changes in production, financial standings, and the role of climate information. A total of 3501 surveys were mailed on October 25, 2002, to ranchs and farms of over 40 acres in the state of Colorado, with 946 usable returned surveys for an effective response rate of ~27%. Analyses of these responses indicate that 30% of the agricultural producers realized the drought in Colorado would significantly impact their production prior to March, 2002, and 55% of the agricultural producers realized this by April, 2002. Over 15% of the responding agricultural producers indicated that they had already started making changes in their production practices before March, 2002, and the number agricultural producers that had started making changes had increased to almost 60% by May, 2002. Almost 50% of the responding agricultural producers indicated that they would have made different production decisions in 2002 if they had known earlier about the drought. About half of these respondents indicated for an accurate drought forecast for Colorado to be useful they would have needed it by January 2002, and less than 10 percent indicated that an accurate drought forecast would have been useful after April, 2002. Other relevant survey questions explored where agricultural producers in Colorado go to obtain weather forecasts, extended forecasts, and seasonal climate outlooks. We will present these results and examine differences in the responses, needs, and pathways for providing information for decision making based on the type of agriculture producer (irrigated farming, dryland farming, and livestock grazing and feeding) and the size of the farm or ranch.