[GRASS-dev] sample vector temporal data

Hi,

as you know we need to decide which data are we going to use for t.vect.* examples. One possibility is to use oceanfront shorelines of North Carolina, we can get this data easily from here:

http://portal.ncdenr.org/web/cm/download-spatial-data-maps-oceanfront

It includes these years, interval data in case of older data, and instance data from the past years:
1849 - 1873, 1925 - 1946, 1933 - 1952, 1940 - 1962, 1970 - 1988, 1997, 1998, 2003, 2004, 2009

The advantage is that the data is public and basically ready to use. If we decide to use it, should we include the entire NC shoreline or just some detail (look at the attachment)?

In addition, we can also create a vector time-series where the geometry is not changing, just the attribute changes (derived from the climate data for example).

Any thoughts on this?

By the way, the climate dataset seems to be fine, we got the confirmation we can use it, it is not subject to the PRISM licence since it was interpolated at NC State Climate Office.

Anna

(attachments)

all.jpg
detail.jpg

Anna,

to put these data into context, the dataset should include also at least one ortho (probably for 2009) and perhaps
also a lidar-based DEM for 2009.
I think that the detail will work better - maybe you can show it on orthophoto or a DEM to make it easier to understand what
the lines mean.

To derive vector time series from the climate data where the geometry is changing just derive isolines for temperature
or precipitation from the existing rasters. If the isolines are noisy we can smooth them out using v.generalize or reinterpolate some data.
This may be a simpler solution because it will use the same data set.

But the shoreline may be more interesting.

Helena

Helena Mitasova
Professor at the Department of Marine,
Earth, and Atmospheric Sciences
and Center for Geospatial Analytics
North Carolina State University
Raleigh, NC 27695-8208
hmitaso@ncsu.edu
http://geospatial.ncsu.edu/osgeorel/
"All electronic mail messages in connection with State business which are sent to or received by this account are subject to the NC Public Records Law and may be disclosed to third parties.”

On Oct 7, 2014, at 9:21 PM, Anna Petrášová wrote:

Hi,

as you know we need to decide which data are we going to use for t.vect.* examples. One possibility is to use oceanfront shorelines of North Carolina, we can get this data easily from here:

http://portal.ncdenr.org/web/cm/download-spatial-data-maps-oceanfront

It includes these years, interval data in case of older data, and instance data from the past years:
1849 - 1873, 1925 - 1946, 1933 - 1952, 1940 - 1962, 1970 - 1988, 1997, 1998, 2003, 2004, 2009

The advantage is that the data is public and basically ready to use. If we decide to use it, should we include the entire NC shoreline or just some detail (look at the attachment)?

In addition, we can also create a vector time-series where the geometry is not changing, just the attribute changes (derived from the climate data for example).

Any thoughts on this?

By the way, the climate dataset seems to be fine, we got the confirmation we can use it, it is not subject to the PRISM licence since it was interpolated at NC State Climate Office.

Anna

<all.jpg><detail.jpg>_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

On 8 October 2014 at 03:21, Anna Petrášová <kratochanna@gmail.com> wrote:

Hi,

Hi,

as you know we need to decide which data are we going to use for t.vect.*
examples. One possibility is to use oceanfront shorelines of North Carolina,
we can get this data easily from here:

http://portal.ncdenr.org/web/cm/download-spatial-data-maps-oceanfront

It includes these years, interval data in case of older data, and instance
data from the past years:
1849 - 1873, 1925 - 1946, 1933 - 1952, 1940 - 1962, 1970 - 1988, 1997, 1998,
2003, 2004, 2009

The advantage is that the data is public and basically ready to use. If we
decide to use it, should we include the entire NC shoreline or just some
detail (look at the attachment)?

In addition, we can also create a vector time-series where the geometry is
not changing, just the attribute changes (derived from the climate data for
example).

Any thoughts on this?

After some weeks without any comments I think we could proceed to
create the dataset.

By the way, the climate dataset seems to be fine, we got the confirmation we
can use it, it is not subject to the PRISM licence since it was interpolated
at NC State Climate Office.

I would like to work on it and on temporal documentation in the next
two days, so what do you think about this plan starting from the
sample data for the temporal workshop [0]?
- remove some years from climate_2000_2012 data, for example keep no
more that 5 years, because right know the location is a little bit
heavy (about 700MB) and for documentation 5 years should be fine.
- create vector for static temporal dataset from towns querying the
rasters in climate_2000_2012 to create something like virtual weather
stations
- add the shoreline ocean data, maybe in a new mapset
- create at least one temporal dataset for each type (strds, str3ds, stvds)
- update the temporal documentation according the new dataset.

what do you think?

Helena in her answer was speaking about orthophoto and DEM for the
shoreline ocean data, are this data already inside the location
(elev_lid792_1m,elev_state_500m,ortho_t792_1m) ?

Anna

[0] http://fatra.cnr.ncsu.edu/temporal-grass-workshop/

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

On 26/11/14 18:10, Luca Delucchi wrote:

On 8 October 2014 at 03:21, Anna Petrášová <kratochanna@gmail.com> wrote:

Hi,

Hi,

as you know we need to decide which data are we going to use for t.vect.*
examples. One possibility is to use oceanfront shorelines of North Carolina,
we can get this data easily from here:

http://portal.ncdenr.org/web/cm/download-spatial-data-maps-oceanfront

It includes these years, interval data in case of older data, and instance
data from the past years:
1849 - 1873, 1925 - 1946, 1933 - 1952, 1940 - 1962, 1970 - 1988, 1997, 1998,
2003, 2004, 2009

The advantage is that the data is public and basically ready to use. If we
decide to use it, should we include the entire NC shoreline or just some
detail (look at the attachment)?

In addition, we can also create a vector time-series where the geometry is
not changing, just the attribute changes (derived from the climate data for
example).

Any thoughts on this?

After some weeks without any comments I think we could proceed to
create the dataset.

By the way, the climate dataset seems to be fine, we got the confirmation we
can use it, it is not subject to the PRISM licence since it was interpolated
at NC State Climate Office.

I would like to work on it and on temporal documentation in the next
two days, so what do you think about this plan starting from the
sample data for the temporal workshop [0]?
- remove some years from climate_2000_2012 data, for example keep no
more that 5 years, because right know the location is a little bit
heavy (about 700MB) and for documentation 5 years should be fine.

+1

- create vector for static temporal dataset from towns querying the
rasters in climate_2000_2012 to create something like virtual weather
stations

Not sure what you mean by static temporal dataset ? Why do we need this in addition to the below shoreline data ?

- add the shoreline ocean data, maybe in a new mapset

+1, including for separate mapset

- create at least one temporal dataset for each type (strds, str3ds, stvds)

+1, but as you say yourself, be careful about mapset size !

- update the temporal documentation according the new dataset.
what do you think?

Sounds good to me.

Thanks a lot for taking care of this !

Moritz

On Wed, Nov 26, 2014 at 12:10 PM, Luca Delucchi <lucadeluge@gmail.com>
wrote:

On 8 October 2014 at 03:21, Anna Petrášová <kratochanna@gmail.com> wrote:
> Hi,
>

Hi,

> as you know we need to decide which data are we going to use for t.vect.*
> examples. One possibility is to use oceanfront shorelines of North
Carolina,
> we can get this data easily from here:
>
> http://portal.ncdenr.org/web/cm/download-spatial-data-maps-oceanfront
>
> It includes these years, interval data in case of older data, and
instance
> data from the past years:
> 1849 - 1873, 1925 - 1946, 1933 - 1952, 1940 - 1962, 1970 - 1988, 1997,
1998,
> 2003, 2004, 2009
>
> The advantage is that the data is public and basically ready to use. If
we
> decide to use it, should we include the entire NC shoreline or just some
> detail (look at the attachment)?
>
> In addition, we can also create a vector time-series where the geometry
is
> not changing, just the attribute changes (derived from the climate data
for
> example).
>
> Any thoughts on this?
>

After some weeks without any comments I think we could proceed to
create the dataset.

> By the way, the climate dataset seems to be fine, we got the
confirmation we
> can use it, it is not subject to the PRISM licence since it was
interpolated
> at NC State Climate Office.
>

I would like to work on it and on temporal documentation in the next
two days, so what do you think about this plan starting from the
sample data for the temporal workshop [0]?
- remove some years from climate_2000_2012 data, for example keep no
more that 5 years, because right know the location is a little bit
heavy (about 700MB) and for documentation 5 years should be fine.

I agree.

- create vector for static temporal dataset from towns querying the
rasters in climate_2000_2012 to create something like virtual weather
stations

yes, but there are no NC towns as points in the standard dataset. You could
use precip_30ynormals@PERMANENT which are the meteorology stations, which
probably doesn't make much sense since but maybe it's still good for the
dataset.

- add the shoreline ocean data, maybe in a new mapset

Helena suggests here to select a smaller, dynamic area (some cape for
example) and also provide a DEM and ortho for that area (elev_lid792 is not
coastal) to get some context. I am not sure about this dataset, if it's
really needed, because I don't know what kind of temporal vector analysis
we could show in the manual. It could be a good dataset to show how to
create animations. If we decide to do it, I can help you with this part.

Another option is to derive contours from the temperature/precipitation
dataset. The advantage is that's easier to prepare.

- create at least one temporal dataset for each type (strds, str3ds, stvds)

Do we have any temporal data for 3d raster? I don't think we necessarily
have to create this dataset.

Anna

- update the temporal documentation according the new dataset.

what do you think?

Helena in her answer was speaking about orthophoto and DEM for the
shoreline ocean data, are this data already inside the location
(elev_lid792_1m,elev_state_500m,ortho_t792_1m) ?

> Anna
>

[0] http://fatra.cnr.ncsu.edu/temporal-grass-workshop/

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

On Nov 26, 2014, at 2:22 PM, Anna Petrášová wrote:

On Wed, Nov 26, 2014 at 12:10 PM, Luca Delucchi <lucadeluge@gmail.com> wrote:
On 8 October 2014 at 03:21, Anna Petrášová <kratochanna@gmail.com> wrote:

> as you know we need to decide which data are we going to use for t.vect.*
> examples. One possibility is to use oceanfront shorelines of North Carolina,
> we can get this data easily from here:
>
> http://portal.ncdenr.org/web/cm/download-spatial-data-maps-oceanfront
>
> It includes these years, interval data in case of older data, and instance
> data from the past years:
> 1849 - 1873, 1925 - 1946, 1933 - 1952, 1940 - 1962, 1970 - 1988, 1997, 1998,
> 2003, 2004, 2009
>
> The advantage is that the data is public and basically ready to use. If we
> decide to use it, should we include the entire NC shoreline or just some
> detail (look at the attachment)?
>
> In addition, we can also create a vector time-series where the geometry is
> not changing, just the attribute changes (derived from the climate data for
> example).
>
> Any thoughts on this?
>

After some weeks without any comments I think we could proceed to
create the dataset.

> By the way, the climate dataset seems to be fine, we got the confirmation we
> can use it, it is not subject to the PRISM licence since it was interpolated
> at NC State Climate Office.
>

I would like to work on it and on temporal documentation in the next
two days, so what do you think about this plan starting from the
sample data for the temporal workshop [0]?
- remove some years from climate_2000_2012 data, for example keep no
more that 5 years, because right know the location is a little bit
heavy (about 700MB) and for documentation 5 years should be fine.

I agree.

- create vector for static temporal dataset from towns querying the
rasters in climate_2000_2012 to create something like virtual weather
stations

yes, but there are no NC towns as points in the standard dataset. You could use precip_30ynormals@PERMANENT which are the meteorology stations, which probably doesn't make much sense since but maybe it's still good for the dataset.

Anna is right - in fact these are the stations which were used to create the rasterized climate time series so there is no need to create
virtual weather stations - we already have the real ones. We can get more complete data for these stations if needed
- all have at least temperature and precipitation on daily basis going back several decades.

- add the shoreline ocean data, maybe in a new mapset

Helena suggests here to select a smaller, dynamic area (some cape for example) and also provide a DEM and ortho for that area (elev_lid792 is not coastal) to get some context. I am not sure about this dataset, if it's really needed, because I don't know what kind of temporal vector analysis we could show in the manual. It could be a good dataset to show how to create animations. If we decide to do it, I can help you with this part.

Another option is to derive contours from the temperature/precipitation dataset. The advantage is that's easier to prepare.

- create at least one temporal dataset for each type (strds, str3ds, stvds)

Do we have any temporal data for 3d raster? I don't think we necessarily have to create this dataset.

we can ask our colleagues for some 3D atmospheric data or the ocean temperature or salinity data,
but it would take us some time to get to this. We also have some good interpolated 3D/4D soil data but they are not public
at this point.

Helena

Anna

- update the temporal documentation according the new dataset.

what do you think?

Helena in her answer was speaking about orthophoto and DEM for the
shoreline ocean data, are this data already inside the location
(elev_lid792_1m,elev_state_500m,ortho_t792_1m) ?

> Anna
>

[0] http://fatra.cnr.ncsu.edu/temporal-grass-workshop/

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

On Wed, Nov 26, 2014 at 8:22 PM, Anna Petrášová <kratochanna@gmail.com> wrote:
...

yes, but there are no NC towns as points in the standard dataset.

There is a (small) geonames layer in the standard NC dataset which we
could add (updated).
But polygons would be better for zonal statistics.

...

Helena suggests here to select a smaller, dynamic area (some cape for
example) and also provide a DEM and ortho for that area (elev_lid792 is not
coastal) to get some context. I am not sure about this dataset, if it's
really needed, because I don't know what kind of temporal vector analysis we
could show in the manual.

Derive contour lines from the LiDAR time series?
Another option is to enrich it with vectorized NLCD time series:

http://www.mrlc.gov/finddata.php
- National Land Cover Database 2011 (NLCD2011)
- National Land Cover Database 2006 (NLCD2006)
- National Land Cover Database 2001 (NLCD2001)
- National Land Cover Dataset 1992 (NLCD1992)

Do we have any temporal data for 3d raster? I don't think we necessarily
have to create this dataset.

3D point soil data would suffice...

Markus

On Wed, Nov 26, 2014 at 3:10 PM, Markus Neteler <neteler@osgeo.org> wrote:

On Wed, Nov 26, 2014 at 8:22 PM, Anna Petrášová <kratochanna@gmail.com>
wrote:
...
> yes, but there are no NC towns as points in the standard dataset.

There is a (small) geonames layer in the standard NC dataset which we
could add (updated).
But polygons would be better for zonal statistics.

I think we need points for t.vect.what.strds, t.vect.observe.strds.

...
> Helena suggests here to select a smaller, dynamic area (some cape for
> example) and also provide a DEM and ortho for that area (elev_lid792 is
not
> coastal) to get some context. I am not sure about this dataset, if it's
> really needed, because I don't know what kind of temporal vector
analysis we
> could show in the manual.

Derive contour lines from the LiDAR time series?
Another option is to enrich it with vectorized NLCD time series:

http://www.mrlc.gov/finddata.php
- National Land Cover Database 2011 (NLCD2011)
- National Land Cover Database 2006 (NLCD2006)
- National Land Cover Database 2001 (NLCD2001)
- National Land Cover Dataset 1992 (NLCD1992)

I remember one student had problem with some of these older datasets. I
haven't tried myself.
Anyway, we have a lot of options... We have to decide for one.

> Do we have any temporal data for 3d raster? I don't think we necessarily
> have to create this dataset.

3D point soil data would suffice...

do we have any? We would have to have a timeseries of points for a str3ds.

Anna

Markus

On Wed, Nov 26, 2014 at 9:31 PM, Anna Petrášová <kratochanna@gmail.com> wrote:

On Wed, Nov 26, 2014 at 3:10 PM, Markus Neteler <neteler@osgeo.org> wrote:

On Wed, Nov 26, 2014 at 8:22 PM, Anna Petrášová <kratochanna@gmail.com>
> Do we have any temporal data for 3d raster? I don't think we necessarily
> have to create this dataset.

3D point soil data would suffice...

do we have any? We would have to have a timeseries of points for a str3ds.

Perhaps from SSURGO Soil Map Coverage

and/or

http://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm

Markus

On 26 November 2014 at 20:22, Anna Petrášová <kratochanna@gmail.com> wrote:

yes, but there are no NC towns as points in the standard dataset. You could
use precip_30ynormals@PERMANENT which are the meteorology stations, which
probably doesn't make much sense since but maybe it's still good for the
dataset.

Ok, I saw them

Helena suggests here to select a smaller, dynamic area (some cape for
example) and also provide a DEM and ortho for that area (elev_lid792 is not
coastal) to get some context. I am not sure about this dataset, if it's
really needed, because I don't know what kind of temporal vector analysis we
could show in the manual. It could be a good dataset to show how to create
animations. If we decide to do it, I can help you with this part.

Another option is to derive contours from the temperature/precipitation
dataset. The advantage is that's easier to prepare.

as you prefer, for me it is the same...

Anna

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

On 26 November 2014 at 21:02, Helena Mitasova <hmitaso@ncsu.edu> wrote:

Anna is right - in fact these are the stations which were used to create the rasterized climate time series so there is no need to create
virtual weather stations - we already have the real ones. We can get more complete data for these stations if needed
- all have at least temperature and precipitation on daily basis going back several decades.

Perfect

we can ask our colleagues for some 3D atmospheric data or the ocean temperature or salinity data,

atmospheric data or the ocean temperature could be really interesting,
could you ask?

Helena

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

On 26 November 2014 at 21:10, Markus Neteler <neteler@osgeo.org> wrote:

But polygons would be better for zonal statistics.

As Anna wrote we need points, but we should have both. Maybe
census_wake2000 or nc_state?

Markus

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

On 26 November 2014 at 21:31, Anna Petrášová <kratochanna@gmail.com> wrote:

Anyway, we have a lot of options... We have to decide for one.

yes, the proposal are:

- shoreline ocean
- contour from precip and/or temp
- contour from LIDAR

so?

do we have any? We would have to have a timeseries of points for a str3ds.

if I understand well we already have, the precip_30ynormals_3d we
could extract layer for each months

Anna

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

On 26 November 2014 at 20:22, Anna Petrášová <kratochanna@gmail.com> wrote:

I agree.

Maybe we should add daily temperature far a year, (they are usefull
for t.rast.accumulate) do you have these data?

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

On Thu, Nov 27, 2014 at 4:21 AM, Luca Delucchi <lucadeluge@gmail.com> wrote:

On 26 November 2014 at 20:22, Anna Petrášová <kratochanna@gmail.com>
wrote:
>
> I agree.
>

Maybe we should add daily temperature far a year, (they are usefull
for t.rast.accumulate) do you have these data?

No, there are only monthly data which can be freely used for the sample
dataset. They have the daily measurements on the stations I assume, but we
would have to interpolate it. But I thought you have daily temperatures in
your LST dataset?

Anna

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org

Hi devs,

In r63363 I submitted an update of space time doc for several modules
using the new temporal dataset.

I'm going to make the new dataset available in the next days to test
and comment it.

Stay tuned :wink:

--
ciao
Luca

http://gis.cri.fmach.it/delucchi/
www.lucadelu.org