Species abundance is often spatially structured such that, within a species’ distribution, abundance peaks at one or more areas and declines from those points. Abundance may also increase or decrease over time, but the spatial structure of temporal changes in abundance has been infrequently examined. Here we use count data from the North American Breeding Bird Survey (BBS) to describe the spatial structure of yearly changes in abundance across the distributions of 135 species of land birds. For each species, we calculated the difference in the logarithms of the number of birds counted from one year to the next at survey routes within their distributions. We assessed the spatial structure of yearly changes in abundance using Moran’s I, a measure of spatial autocorrelation. For comparison, we also calculated Moran’s I for abundance, i.e., the logarithm of the number of birds counted on survey routes in each year. As expected, abundance was positively spatially autocorrelated (i.e. closer routes were more similar than expected by chance) at distances of up to a few hundred kilometers for most species. In contrast, changes in abundance showed little to no spatial autocorrelation. Resident species exhibited greater spatial structure in abundance than migrant species; however, the two groups did not differ in the degree of spatial structure in change in abundance. Variation in abundance over multiple years was mostly unrelated to the distance from the abundance-weighted center of species’ distributions. Temporal changes in abundance can occur at fine spatial scales for many species, but understanding the causes of such changes is challenging. These results fill a gap in the ecological literature and may have important implications for conservation planning.