There are commercially available resin strips designed to be embedded in soil, retrieved and extracted. The process is conceptually nice because it nutrient binding is diffusion controlled, but selectivity varies between ions, so as loading proceeds some ions can displace others. That is you need to try different contact times. The strips only work with good contact and while the soil moisture is sufficient. Spatial variation is also very large at this scale so you need a large number of such devices to represent an area.
What do you want to monitor and why over so large an area? How uniform is it?
Conventional sampling and analysis approaches are designed to overcome many of the above difficulties but that's not to say they are perfect.