This forum is shutting down! Please post new discussions at

SDR model

edited July 2017 in Freshwater Models
I am having some trouble understanding why my SDR model shows a bigger sediment retention by the forest in a scenario with less forest. 
I am running 3 situations, the first one is the actual land cover of my watershed, the second is a scenario increasing the forest as much as possible in the Brazilian Law, and the third one I am increasing the forest but not so much as the second scenario.
Someone can help me?

Use and land cover Retenção de sedimentos
Actual Scenario Cenário I  Cenário II
humid area 46.928 64.085 68.253
urban area 97.726 98.518 97.931
sugar cane 138.080 138.027 138.128
water bodies 16.627 15.071 15.599
Citrus 161.911 162.115 161.981
roads 154.512 152.424 153.770
forest 474.092 417.672 426.727
Pasture 167.061 171.649 170.859
Total 1256.938 1219.562 1233.249
Post edited by victor_rosario on


  • swolnyswolny Member, NatCap Staff
    edited July 2017
    Hi Victor -

    First, a reminder that the sediment retention results should not be used as absolute values, only a relative index.

    How did you get the values in your table? I'm assuming that you summed up the per-pixel sediment retention within each land cover type - yes? If not, let me know how you did get these values. 

    If you look in the Sediment Retention Services section of the SDR User Guide, you'll see how the retention index is computed: It takes the difference between potential erosion from the current landscape, and a landscape that is all bare soil, and then weights that difference by the SDR factor. It is the weighting by SDR that is probably causing the reduction in retention that you see. Yes, if you only compare the potential erosion between more forest and less forest, you'd expect an increase in retention. But we also consider what's going on upslope and downslope in the SDR factor. If there's more forest upslope of each pixel, then the SDR factor should be lower, since there's less erosion being produced by the increased forest. If there's more forest downslope of each pixel, then the SDR factor could also be lower, since there's more erosion being retained between each pixel and the stream. You could check this by looking at the InVEST output intermediate_outputs/SDR_factor.tif and seeing how it compares between scenarios. (To do this, you'd need to rename SDR_factor.tif before rerunning, or the model will overwrite the file.)

    ~ Stacie

    Post edited by swolny on
Sign In or Register to comment.