 |
| Titel |
Determination the optimum number and positions of monitoring stations for proper spatial modeling of mean PM10 concentration in Berlin |
| VerfasserIn |
Hamid Taheri Shahraiyni, Sahar Sodoudi, Ulrich Cubasch |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250097454
|
| Publikation (Nr.) |
 EGU/EGU2014-13042.pdf |
|
|
|
|
|
| Zusammenfassung |
| PM10 concentration in Berlin has exceeded from EU limit, therefore the appropriate
spatial distribution of PM10 concentration is a prerequisite for management and
control of PM10 concentration in Berlin. The key question of this study is “How
many PM10 monitoring stations must be installed in Berlin for appropriate spatial
distribution modeling of PM10 concentration and where do they have to be installed?” In
this study, a geostatistical calculation has been utilized to answer this question.
The optimum number of monitoring stations and their positions are determined
by minimization of estimation variance. The experimental variogram values of
mean PM10 concentration were calculated using the data of 13 existing stations in
Berlin and several variogram models were fitted to the experimental variogram
data. The results demonstrated that circular model is the best variogram model
for mean PM10 concentration and consequently, a circular variogram model was
developed. The Berlin urban area was gridded to 500x500m pixels and for each pixel the
estimation variance was calculated using the mean PM10 concentration of monitoring
stations and the developed circular variogram model. Then, mean estimation variance
(ÏăE2) of Berlin was calculated by averaging of the estimation variance of whole
pixels.
By adding one virtual station in Berlin, the optimum position for this station
was determined using an iterative optimization technique and with the object of
minimization of mean estimation variance (ÏăE2). Hence, the position of added station was
changed and ÏăE2 was calculated iteratively. The best position for added station is the
position that minimizes ÏăE2 value. Using this method, some stations were added
virtually in Berlin one by one and ÏăE2 value was decreased continuously. This
iterative technique was performed until the amount of decrease of ÏăE2 per increase of
one virtual station goes toward zero (|-Ïă2E|
|-n| < 0.04 ). The relation between the
total number of stations (n) (real and virtual) and ÏăE2 was calculated and a power
function was determined [ÏăE2(n) = 10.67n-0.44(R2 = 0.99)]. According to this
power function, |-Ïă2|
-|-En| is less than 0.04, when n -¤ 28. Thus 15 further stations
should be added in the suggested positions. The presented approach not only is a
reliable method for determination of optimum number and positions of air pollution
monitoring stations, but also it is applicable for all spatial correlated data such as
some of meteorological, land surface, water bodies, groundwater and underground
data. |
|
|
|
|
|
|