Spatial dependence means that a variable’s values at different locations are related to or affected by each other depending on their distances. The closer the distance, the more similar the values of the variable, and conversely.

This is according to Tobler's first law of geography, Everything is related to everything else, but near things are more related than distant things. It reflects the fact that the characteristics of the observations are affected by their spatial arrangement, and the values of observations are related to each other through distance.

Spatial dependence is also called or measured as spatial autocorrelation. A typical example of spatial dependence is house prices in a district. The more expensive houses are closer to each other, and the price of a house sold would affect the selling prices of its neighbors. Spatial dependence is one major type of spatial effects a spatial machine learning model needs to take into consideration when it exists.