Parameter setting

Hello, I have some questions, can you help me to solve my problems.
if I want to evaluate the ecosystem services of coastal wetland, and the training data  do not have the reference data of coastal wetland, for example, the land use is Suaeda glauca Bge and Spartina anglica Hubbard, how can I set the parameter setting of Kc, root_depth, usle_c, usle_p,sedret_eff,load_n,eff_n,load_p,eff_p and so on.

Comments

  • swolnyswolny Member, NatCap Staff
    Hi @XiaopingSun -

    First, it's important to know that you should not use the training data as final inputs for your analysis. You can use it as a first test of the models in your area, but the sources of the training data are not documented, and in some cases have been entirely made up, so we strongly recommend against using the training data for anything other than training.

    When we do projects, we do a literature search to find the biophysical table values for our land use types. Try to find data from analyses that were done as close to your area of interest as possible, or on a similar type of land cover, or in a similar ecoregion. If you cannot find local values, then there are often global sources of these values available. In the InVEST User Guide, most models have a Data Sources section, that lists some of these global data sources.

    There are also databases of literature sources for Sediment and Nutrient that we have collected, and you can also use this to find values that might be appropriate for your land cover types. You'll find a link to the database at the bottom of this web page

    ~ Stacie


     
  • <First, it's important to know that you should not use the training data as final inputs for your analysis. You can use it as a first test of the models in your area, but the sources of the training data are not documented, and in some cases have been entirely made up, so we strongly recommend against using the training data for anything other than training.>

     This is really good to know !

Sign In or Register to comment.