Core tests have demonstrated that decreasing the salinity of injection water can increase oil recovery. Although recovery is enhanced by simply decreasing salt content, optimising injection water salinty would offer a clear economic advantage for several reasons. Too low salinity risks swelling of the clays which would lead to permanent reservoir damage but evidence of effectiveness at moderate salinity would offer the opportunity to dispose of produced water. The goal is to define boundary conditions so injection water salinity is high enough to prevent reservoir damage and low enough to induce the low salinity effect while keeping costs and operational requirements at a minimum. Traditional core plug testing for optimising conditions has some limitations. Each test requires a fresh sample, core testing requires sophisticated and expensive equipment, and reliable core test data requires several months because cores must be cleaned, restored and aged before the tests can begin. It is also difficult to compare data from one core with results from another because no two cores are identical, making it difficult to distinguish between effects resulting from different conditions and effects resulting from different cores. Gathering statistics is limited by the time required for each test and the fact that core material is in short supply. Thus, our aim was to explore the possibility of a cheaper, faster alternative. We developed a method that uses atomic force microscopy (AFM) to investigate the relationship between the wettability of pore surfaces and water salinity. We functionalise AFM tips with organic molecules and use them to represent tiny oil droplets of nonpolar or polar molecules and we use sand grains removed from core plugs to represent the pore walls in sandstone. We bring our "oil" close to the sand grain surface and measure the work of adhesion between the tip and the surface. Repeated "feeling" the surface with the tip produces data that can be converted to maps of adhesion and contact angle. Adhesion work is proportional to wettability and is directly correlated with the salinity of the fluid in contact with the tip and the particle surface. From our measurements, the threshold values for the onset of the low salinity response is 5,000 to 8,000 ppm, which benchmarks remarkably well with observations from core plug tests. Changing either the type of "oil" on our probe or the character of the grain surface both affect the adhesion response. From a mechanistic perspective, the correlation between salinity and adhesion provides evidence for the role of electrical double layer expansion in the low salinity response; expansion of the double layer decreases oil wettability. Because AFM experiments can be done relatively quickly on very little material, it gives the possibility of testing salinity response on samples from throughout a reservoir and for gathering statistics. Our approach provides a range of data that can be used to screen core plug testing conditions and to provide extra data that would be too time consuming or too expensive using traditional methods alone.