Model runs from 1981-1-1 to 2099-12-31
The following variables are arguments of the function snow_model() which is called by the EMA-workbench.
Climate eXternalities/uncertainties:
xRCP : (from 0.51 to 3.49) Represents the concentration pathway (i.e. the climate scenario), either RCP2.6, RCP4.5 or RCP8.5
Is rounded to a value 1, 2 or 3?
xClimateModel : (from 0 to 1) Represents the climate model
Model eXternalities/uncertainties: ?What do these represent?
X2fM : (from 1.01 to 1.61)
X3iPt : (from 900 to 1100)
X5temp : (from 0 to 6.0)
X6tempArt : (from -2.0 to -1.0) ??? Temperature at which we start making articial snow?
Constants:
xCostDay : (constant)
xRevenueDay : (constant)
Xfactor1 : (constant)
X4rSnow : (constant)
Policy Levers:
s1SnowThershold : ?? The point where we start making artifical snow ??
xGoodDays : Threshold for number of good snow conditions for making a profit?
Returns:
y : AveragesumRows (Scalar) ~ something with artificial snow making (from 20 to 120)
y1 : climateModel (Scalar) ~ the (input???) climate model? (from 0 to 70)
y3 : sumRows (Array) ~ the number of tipping points per year? (from 2500 to 13,000)
y4 : AveragesumRowsArtSnow (Scalar) ~ summary of days with artifical snow making? (from 5000 to 10,000)
y4 : AveragesumRowsProfit (Scalar) ~ average of days in which profit is made? (from 0 to -150)
y6 : AveragereportMatrix (Scalar) ~ ?? (from 0.1 to 0.8)
y: Seasonal average (1981-2100) of days with Good Snow Conditions for Skiing (snow depth > 200 - 300 mm)
y1: The Climate model which has been used (Should not be presented in the outcomes)
y3: Seasonal (1981-2100) number of days with Good Snow Conditions for Skiing (snow depth > 200 - 300 mm)
y4: Seasonal average (1981-2100) of number of days with the possibility of Artificial Snow Making (ave TEMP < -2 to -1)
y5: Relative seasonal average (1981-2100) of Financial balance of Ski resort
y6: Seasonal average (1981-2100) Likelihood of survival of a Ski resort from tipping points
y7: Number of tipping points (1981-2100)
These are for Hoch-Ybrig
grouped by policy gives 40 results, how are these composed?
Metrics y and y3 have a perfect linear relation, so I would guess on of them is superfluous
Find subregions of the input space with Metric values of interest.
Apparantly searching for outcomes['y'] < 50 -> TRUE
Assume: TRUE means tipping point.
In the pairs scatter plot with the boxes I can see: clear box of outcomes around X5temp
Feature scoring diagram (~block 45). This is sort of a regression test between input (X,L) parameters and outcome metrics (M)
As we also saw in the SD: the FS shows that Metric y is very sensitive to changes in the X5te
Metric y1 is very sensitive to the chosen RCP.
Metrics y3, y5 and y6 behave similar as y
Metric y4 is sensitive to xRCP and X6tempArt.
It seems that the policy Levers do not have any serious impact on the results: policy, x1SnowThreshold and xGoodDays have little effect on outcome parameters.
Questions: Why implement xRCP as RealParameter, in practice it can only have 3 discrete values? Maybe easier with the IntegerParameter (0,1,2) or CategoricalParameter ('RCP26','RCP45','RCP85') ?