Weather is a critical factor for agricultural production. How growers perceive weather and changes in weather trends will determine land use and other agricultural production practices in a region. This study presents a spatial-temporal analysis of localized weather perceptions of agricultural growers from a Farm Survey in the eastern portions of North Dakota and South Dakota. Past (2004 to 2014) and future (2014 to 2024) perceptions of four weather indictors were studied: temperature, precipitation, drought frequency and flood frequency. We found that weather perceptions exhibit positive spatial autocorrelation and respondents with similar beliefs are located in localized clusters. About half of the respondents perceived temperature as being the same as the last 10 years (2004-2014) and staying about the same in the next 10 years (2014-2024). Consistent with scientific projection of climate change trends in the region, more respondents perceived that the climate would change from being cooler and wetter to warmer and dryer rather than the other way around in the future 10 years. Further, beliefs about past temperature and droughts were inconsistent with recorded actual weather data. We link land use change with changes in weather to find that regression-based inference would differ whether actual weather data or perceived weather are included as regressors. Behavioral rationales for our statistical findings are provided along with policy implications.