I am conducting a long-term simulation study on how ecosystem services change according to land use/land cover (LULC) in a specific region.
Evapotranspiration was computed using equations, and precipitation was interpolated from observation stations across the study area using IDW (inverse distance weighting).
When I obtained my results, I was initially confident that forests would have higher AWY values than bare land, cropland, and urban areas.
However, upon reviewing the results, I found that my assumption was wrong.
Generally, we expect forests to have a higher water yield than bare land, cropland, and urban areas.
Given these findings, I am having difficulty interpreting the results.
Are these results because forests have a higher KC than other land types?
If so, how should this be interpreted?
I’d be grateful for any advice
*Starting from 1 is the urban area, 2 is the farlmalnd, 3 is the forest, 5 is the abandoned area
It is usually the case in the AWY model that forest produces lower water yield, and yes you’re correct that it’s because forests tend to evapotranspire a lot of water, so have a higher Kc value.
This feels counter-intuitive, since we want forests to be the best at providing all sorts of ecosystem services. We will often create a reforestation scenario and find that there is less annual water yield from the model. In reality, forests can provide a lot of flow-related services, slowing down water to reduce flooding, helping it go into the ground to increase groundwater recharge and baseflow in the dry season, etc. But the AWY model is very simple and does not capture any of these real-world processes.
The Seasonal Water Yield model provides a little more detail, differentiating quick flow (which runs off quickly and can cause flooding), local recharge (where water goes into the ground) and baseflow (which makes it to streams later, perhaps in the dry season when it’s needed most). This model is still very simple and has a lot of limitations, but it might show more of the forest benefits. If neither of these provide the level of detail that you need, then you’ll need to try a different model that is more complex.
