[R-sig-Geo] spRbind for spplot

[Paul Hiemstra]

On 08/18/2010 11:03 AM, Klaus Vormoor wrote:
> Dear Paul,
>
> thanks for your answer. I guess, your second advice is pricipially 
> that what I want. But so far it did not work. I try a more precise 
> explaination:
>
> For an area I have in total 2200 catchments (nor = readOGR(shp)). 100 
> of them have observations for five days (rtopObj3$observations). I am 
> only interested in the attributes for one day ("QNORM_12_4"). 
> rtopObj3$observations holds spatial information only for these 100 
> catchments. rtopObj3$predicions holds spatial information only for the 
> remaining 2100 catchments with predictions 
> (rtopObj3$predictions$var1.pred).
>
> When I try
>> plot_poly = rtopObj3$observations
>> plot_poly$predictions = rtopObj3$predictions$var1.pred
>> spplot(plot_poly, c("observations", "predictions"))
> it does not work since the replacement has 2100 rows and the data has 
> 100.
>
> I would like to create a new object (e.g. plot_poly) which holds the 
> spatial information of all catchment-polygons (nor) as well as the 
> attributes of both observations$QNORM_12_4 and predictions$var1.pred 
> to the corresponding areas. Therefore, I thought spRbind might be a 
> propper solution.
>
> I hope this is more specific and not even more cryptical...
>
> Best regards,
> Klaus
>
>
>
>
> On Tue, 17 Aug 2010 20:21:43 +0200
>  Paul Hiemstra <p.hiemstra at geo.uu.nl> wrote:
>> Dear Klaus,
>>
>> You need to be more specific as to what you want exactly. From what I 
>> understand both your runoff and predictions are attributes that you 
>> want to use to fill the polygons. My first idea would be to make to 
>> plots next to each other, which is quite simple in spplot:
>>
>> spplot(data, c("attribute1", "attribute2"))
>>
>> If your attributes are not part of the same polygons set you can add 
>> one as a new attribute to the other:
>>
>> plot_poly = rtopObj3$observations
>> plot_poly$predictions = rtopObj3$predictions$var1.pred
>> spplot(plot_poly, c("observations", "predictions"))
>>
>> Is this satisfactory for you, and what do you mean by one map showing 
>> both? To show the results in one figure a scatterplot of observerd vs 
>> modeled is also a good idea.
>>
>> regards,
>> Paul
>>
>> On 08/17/2010 03:06 PM, Klaus Vormoor wrote:
>>> Dear all,
>>>
>>> I have an object (rtopObj3) including the SpatialPolygonDataFrames with
>>> runoff observations ($observations, "QNORM_12_4") for 100 gauged 
>>> catchments
>>> and with predictions ($predictions, "var1.pred") for more than ungauged
>>> 2,000 catchments. So far, I only can plot either the observations or 
>>> the
>>> predictions individually:
>>>
>>>> spplot(rtopObj3$observations, "QNORM_12_4", col.regions = 
>>>> bpy.colors())
>>>> spplot(rtopObj3$predictions, "var1.pred", col.regions = bpy.colors())
>>> I want to create one map showing both. I tried the sp.layout argument
>>> without success. But since I also want to have both information 
>>> combined for
>>> writing a shp-file, it may be more reasonable to use spRbind and do 
>>> the plot
>>> later on. I unsuccessfully tried:
>>>
>>>> observation = rtopObj3$observations
>>>> prediction = rtopObj3$predictions
You need to make a subset of the prediction that includes only the 100 
observed catchments. Maybe even better, create a prediction object of 
only the 100 observed catchments.

cheers,
Paul
>>>> rtopObj4<- spRbind(observation, prediction)
>>> Error in rbind(deparse.level, ...) :
>>>    numbers of columns of arguments do not match
>>>
>>> Does anyone know a way out?
>>> Thanks for every advice,
>>>
>>> Klaus
>>
>>
>> -- 
>> Drs. Paul Hiemstra
>> Department of Physical Geography
>> Faculty of Geosciences
>> University of Utrecht
>> Heidelberglaan 2
>> P.O. Box 80.115
>> 3508 TC Utrecht
>> Phone:  +3130 253 5773
>> http://intamap.geo.uu.nl/~paul
>> http://nl.linkedin.com/pub/paul-hiemstra/20/30b/770
>


-- 
Drs. Paul Hiemstra
Department of Physical Geography
Faculty of Geosciences
University of Utrecht
Heidelberglaan 2
P.O. Box 80.115
3508 TC Utrecht
Phone:  +3130 253 5773
http://intamap.geo.uu.nl/~paul
http://nl.linkedin.com/pub/paul-hiemstra/20/30b/770