We compile panel data on salinity and crop choice for over 13,000 parcels in California's Sacramento-San Joaquin Delta from 2009 to 2016. We use parcel fixed effects in a multinomial logit model to predict crop choice as a function of irrigation water salinity, allowing for unobserved parcel-level heterogeneity that causes spatial variability in needs for salinity management. This contrasts with previous work, where measures of salinity capture the impact of both variation in irrigation water salinity and underlying soil/parcel conditions. We find increased odds of planting both low-value, tolerant crops and certain high-value, salt-sensitive crops relative to orchards. Implications for Delta revenues are discussed.