Hello, community and software support! My name is Victor and I work for the Environmental Institute of Rio de Janeiro, Brazil. We are running both SDR and SWY to test their potential in forest recovery scenarios. SDR has prooven to be most effective. However we have encountered some limitations regarding SWY. For exemple:
Obtaining monthly Kc values for natural or anthropic vegetation like forests or fallow/pastures is demanding a great effort of calibration, without a precise level of certainty. From the work of Allen et al. 1998, can we consider to use average Kc values for pastures, since they are not cut during the year (abandoned pastures)?
Does SWY consider slope from the DEM when calculating water balance, quick flow and base flow?
Hi, I can only answer #2 which is no. Iām guessing you might be thinking high slope would have less contribution to water balance or something? My understanding is the time periods are long enough that is not necessary to consider. Also if you havenāt seen it, the full model for SWY is described here: http://releases.naturalcapitalproject.org/invest-userguide/latest/seasonal_water_yield.html
For #1, I agree that gathering all of those Kc values is very time-consuming and generally imprecise. So we use the best values that we can find, that seem to represent the system well. Iām not quite sure what you mean by āaverageā Kc values for pastures, but if this means assigning the same Kc value for all months of the year, that seems fine if you know that the pastures donāt change much during the year. Iāve often used the same Kc value all year, for example tropical evergreen forests, where there isnāt likely to be much change (and certainly no published change) in Kc during the year. Itās more important (and sometimes easier) to find differing Kc values for agriculture and similar land cover types that definitely change over the year.
Thank you, Stacy! Regarding average Kc, thatās exactly what I meant. I was affraid that using the same values for all months could mask or alter the final result in some way. Indeed, itās been really hard finding literature about Kc for natural vegetated areas, or pastures that are not submited to seasonal cutting or irrigation (my case, specifically). We tried man alternatives, including estimating Kc via remote sensing. May I ask what value did you use for tropical evergreen forests?
Thank you! Yes, iām thinking that high slope areas will contribute less to water recharge or base flow. Iām running the model for a very humid tropical basin, with significant variation in slope and altitude, so the time interval of one year would be significant, I would think. Indeed, it is described in the manual that the model does not consider slope for both quick flow and base flow.
Although this thread was 10 months ago, I find your discussions really helpful. Iām also doing the same process here in South Carolina (SDR and SWY) to value the ecosystem services and as an input to a Marxan software eventually.
But to the Kc point, would it make sense to use the Kc of agriculture based on the cropping season? I used the cropping data layer as landcover since it has detailed crop cover and searched for the Kc of all crops. Then apply the Kc only on months when the crops are in season. For off-season, I used the Kc for barren to simulate the idle cropland period. Would this approach make sense to capture the monthly variation in SWY?
Yes, it definitely makes sense to use the Kc of agriculture based on seasonal variation for this seasonal model. Youāre fortunate to have both detailed crop information and Kc values for them. We have also used the Kc value for bare ground during the fallow season, which of course assumes no cover cropping or other soil management practice is used.
Thanks for the response.
This would allow us now to assess the difference if a simulated cover crop or other practices are implemented, versus just the bare ground for ecosystem service change.
Thank you so much. Your response is very helpful.
Hello there,
Iām here for the answer on question 1. Iām looking for Kc values of forests in the Pantanal region (Brazil, Bolivia, Paraguay). However, few sources that I have found give Kc values for natural forests. The best source I found thus far is (Silva et al, 2018). However, this source appoints Kc values of 1.11-1.25 to natural forests and 0.59-0.88 to pastures. Instinctively, forests should have lower Kc values as they enhance water infiltration, right?
The output of the model looks weird and there seems to be too much evapotranspiration relative to precipitation, causing no local recharge and baseflow whatsoever except a tiny bit on pastures. But nothing in the forested areas. Thatās why I think the issue is in the Kc.
This is my local recharge map. Fully purple are the forest areas (L value: 0), those weirdly squared areas are pastures (L value:0-5).
Since I am new to the SWY model and all of its parameters, could someone clarify the meaning of Kc to me and if the values provided by (Silva et al, 2018) make sense? Thanks!
Silva, C., Lilla Manzione, R., & Albuquerque Filho, J. L. (2018). Large-scale spatial modeling of crop coefficient and biomass production in agroecosystems in southeast Brazil. Horticulturae, 4(4), 44.
It is often the case that Kc is higher for forest. Itās a measure of how much water that particular vegetation type transpires, so larger trees, possibly with understory plants, are likely to take up and transpire more water. The ET0 input quantifies a baseline/reference ET, and itās combined with Kc to calculate the actual ET for that particular vegetation type. (You can read more about that in other posts, and with a web search).
Itās also generally the case that forests are good at infiltration, and slowing down runoff, leading to enhanced baseflow in the dry season. So we often will see lower overall water yield from forests, but it is spread out over a longer time period, which might be important in places with distinct wet and dry seasons. While SWY does provide quickflow results on a monthly basis, it unfortunately doesnāt provide monthly baseflow, so we canāt see the tradeoff directly with this model.
There is some complexity to the calculation of local recharge, but one way to sanity-check the results is by looking at the actual evapotranspiration (AET) output relative to precipitation, corresponding to the Local Recharge equations in the User Guide.
Thank you for the elaborate and clear response Stacie, that helps me with understanding the outputs better. The AET and input ET0 values are much higher than the input P values, leading to B and L values of 0 (except for the agricultural areas because they have a lower Kc I guess). It does feel a bit weird because I always thought that in a normal situation forests should enhance water infiltration. So in the model, the advantage that a forest has, relative to agricultural areas, in terms of enhancing infiltration are only represented by the input Cn values (forest>agri in Cn) and visible in the output QF (forest<agri in QF).
Quick side question: What does the output file āSiā in intermediate_outputs include? I cannot seem to find it in the SWY model guide.
