distribution and parameter estimation method
The Generalized Extreme Value (GEV) distribution (Jenkinson, 1955) was adopted to model the
flood frequency distribution from the AMF series:
Qi(D;T ) =
ei +
ai
ki
(1 [ ln(1 1=T )]ki) if ki 6= 0
ei ailn( ln(1 1=T )) if ki = 0
(7)
where ei is the location parameter, ai is the scale parameter and ki is the shape parameter.
The method of probability weighted
/media/vedurstofan/utgafa/skyrslur/2015/VI_2015_007.pdf
model errors discussed in the previous section, a model time-series
of 2-m temperature or 10-m wind speed, Mi(t), interpolated to the i-th station location, can be
linearly transformed such that the mean square error compared with the local station time-series
is minimised. Generally, the corrected time-series is then given by
M˜i(t) = ai Mi(t)+bi ; (2)
where at each station location, the correction
/media/vedurstofan/utgafa/skyrslur/2014/VI_2014_005.pdf
-out group, assuming the initiative is at the public
side.
Red: inside transport system; blue: direct impact on size & quality of demand for road vehicle
movements; grey: auxiliary services that strongly interact with effects of CC.
Various possible effects of climate change on road infrastructure and its users
The expected effects of a changing climate in Nordic countries imply among others
/media/vedurstofan/NONAM_1st_workshop_summary_v3.pdf
it will be one of the main aspects
of this paper.
2.3. Properties
The elements of a Fuzzy Cognitive Map are as follows:
Concepts: C1, C2, . . ., Cn. These represent the drivers and
constraints that are considered of importance to the issue under
consideration.
State vector: A = (a1, a2, . . ., an), where ai denotes the state of the
node Ci. The state vector represents the value of the concepts
/media/loftslag/Kok_JGEC658_2009.pdf
[Q(t);Q(t 1);Q(t 2)]
E[Q(ui);Q(ui 1);Q(ui 2)]
, ai =
Qb(t)
Qb(ui)
, bi =
Q(t) Qb(t)
Q(ui) Qb(ui)
,
and Qb(t) and Qb(ui) are baseflows calculated using the UKIH baseflow separation method
(Piggott et al., 2005). All rescaling coefficients were limited to a minimum value of 0.25 and a
maximum value of 5.
4.5 Deterministic prediction
A deterministic forecast was derived from the ensemble by taking
/media/vedurstofan/utgafa/skyrslur/2013/VI_2013_008.pdf