Working with the HWSD and the GLCF

[Mat Bond]

I am having trouble separating out the different variables in the
Harmonised World Soil Database. When I use the data in Maxent it
groups them all as one layer, and I get the same problem when opening
the files in Diva. Any ideas on how to do this?

Also does anyone know of other Maxent compatible datasets then
WorldClim available, I find the WorldClim layers useful but heavily
inter-correlated and would like to look at NDVI and Soil salinity, in
addition to any other major factors influencing plant distribution.

[Peter Wilson]

Mat,

The HWSD contains a large number of soil attributes, including topsoil and subsoil sodium exchange capacity (related to salinity). Which one(s) are you trying to use? Many attributes are coded into a few classes and may need recoding to be used in MaxEnt or a GIS. Also, the classes are based on agronomically relevant tolerance limits.

Also, all climate datasets are highly inter-correlated (i.e. have high levels of collinearity). This is particularly true for derived data like the Bioclim variables commonly used in species distribution modelling. The question of how high levels of collinearity affect MaxEnt is a matter of debate (see many posts to this list). Remember that MaxEnt does not need to invert the covariance or correlation matrix so, unlike a generalized linear model, you can still get a model even when very high levels of collinearity exist in the data. How stable and interpretable the model is is still a matter of debate.

An alternate global climate dataset is available on-line from the Climate Research Unit (CRU), University of East Anglia. It is at 10 arc minute resolution and requires some pre-processing to import it into a GIS grid format as well as re-gridding to finer grids if that is needed. You will also have to generate the Bioclim variables if you need them.

Peter Wilson

==========================================
Dr Peter D. Wilson
Postdoctoral Research Fellow
Department of Biological Sciences
Macquarie University, New South Wales, Australia 2109
Email: pwi...@bio.mq.edu.au
>>> Mat Bond <matj...@hotmail.co.uk> 09/01/09 3:17 AM >>>

[michelle g]

Hi Mat
This is a copy of information from my metadata file on how I managed
to get different raster layers for the different data available in the
HWSD data set. (Hence all the file names in the brackets.) I'm not
quite sure from your description if it is what you need. I have
managed to get the soil data working in Maxent, R and ArcGIS by
following the steps below. I hope it helps.
Regards,
Michelle

Michelle Greve

Aarhus University
Department of Biological Science
Ny Munkegade Bygning 1540
DK-8000 Aarhus C
Denmark


Methods used:
The original .bil raster file was exported to a grid format
(hwsd_raster). The Access database was used to produce a query
containing the MU_Global, SU_Symbol and SU_Code columns. (The other
FAO parameters, such as SU_SYM74, which is the original, more detailed
FAO classification, were not available across the whole world or even
the whole of Africa, and were thus not used to construct the soil
map.) There are 36 different soil classes represented by the
‘SU_Symbol’, and SU_Code is a numerical code for the major soil group
(FAO90), used for the HWSD. This query was exported to a .dbf file
(HWSD_SMU_query.dbf), which was then joined to the original grid file
using the ‘Join’ function in ArcGIS. The join was conducted using the
‘Value’ column of hwsd_raster (which is a unique number for each grid
cell) and the MU_Global column of HWSD_SMU_query.dbf. The grid was
then exported, so that it contained the values from the query table as
attributes in the attribute table of the grid. This new grid with the
attributes was called hwsd_grd_join.
In order to change the ‘Value’ of the raster dataset from the original
unique raster cell values to soil characteristics, the Lookup function
in Spatial Analyst Toolbox was used. The above-mentioned procedure can
only be done using numerical characters, therefore SU_Code was used as
the lookup field. The raster that was produced was saved as
soilcode_hwsd. This file was a floating point table, and was therefore
converted to integer (as the SU_Symbols are categorical integers) in
the raster calculator, using the calculation int([soilcode_hwsd]). The
resulting grid was named hwsd_soil.
Therefore hwsd_soil is the final dataset (other files might be useful
if somebody wanted to extract other information from the dataset – I
have therefore placed the files that were used in the processing of
the data in the folder Edited/Files used in processing).

On Aug 31, 11:16 pm, "Peter Wilson" <pwil...@bio.mq.edu.au> wrote:
> Mat,
>
> The HWSD contains a large number of soil attributes, including topsoil and subsoil sodium exchange capacity (related to salinity). Which one(s) are you trying to use? Many attributes are coded into a few classes and may need recoding to be used in MaxEnt or a GIS. Also, the classes are based on agronomically relevant tolerance limits.
>
> Also, all climate datasets are highly inter-correlated (i.e. have high levels of collinearity). This is particularly true for derived data like the Bioclim variables commonly used in species distribution modelling. The question of how high levels of collinearity affect MaxEnt is a matter of debate (see many posts to this list). Remember that MaxEnt does not need to invert the covariance or correlation matrix so, unlike a generalized linear model, you can still get a model even when very high levels of collinearity exist in the data. How stable and interpretable the model is is still a matter of debate.
>
> An alternate global climate dataset is available on-line from the Climate Research Unit (CRU), University of East Anglia. It is at 10 arc minute resolution and requires some pre-processing to import it into a GIS grid format as well as re-gridding to finer grids if that is needed. You will also have to generate the Bioclim variables if you need them.
>
> Peter Wilson
>
> ==========================================
> Dr Peter D. Wilson
> Postdoctoral Research Fellow
> Department of Biological Sciences
> Macquarie University, New South Wales, Australia 2109

> Email: pwil...@bio.mq.edu.au
>
> >>> Mat Bond <matjb...@hotmail.co.uk> 09/01/09 3:17 AM >>>

[Mat Bond]

Dr Peter Wilson,

Thanks,

The topsoil and subsoil sodium exchange capacity both sound ideal, the
HWSD has variables labelled simply as topsoil salinity and subsoil
salinity are these the same variables?

I am basically looking for anything I can use as a proxy indicator of
the salinity of soil and water sources available to plants growing in
the area.

Cheers for the help,

Mat

On Aug 31, 10:16 pm, "Peter Wilson" <pwil...@bio.mq.edu.au> wrote:
> Mat,
>
> The HWSD contains a large number of soil attributes, including topsoil and subsoil sodium exchange capacity (related to salinity). Which one(s) are you trying to use? Many attributes are coded into a few classes and may need recoding to be used in MaxEnt or a GIS. Also, the classes are based on agronomically relevant tolerance limits.
>
> Also, all climate datasets are highly inter-correlated (i.e. have high levels of collinearity). This is particularly true for derived data like the Bioclim variables commonly used in species distribution modelling. The question of how high levels of collinearity affect MaxEnt is a matter of debate (see many posts to this list). Remember that MaxEnt does not need to invert the covariance or correlation matrix so, unlike a generalized linear model, you can still get a model even when very high levels of collinearity exist in the data. How stable and interpretable the model is is still a matter of debate.
>
> An alternate global climate dataset is available on-line from the Climate Research Unit (CRU), University of East Anglia. It is at 10 arc minute resolution and requires some pre-processing to import it into a GIS grid format as well as re-gridding to finer grids if that is needed. You will also have to generate the Bioclim variables if you need them.
>
> Peter Wilson
>
> ==========================================
> Dr Peter D. Wilson
> Postdoctoral Research Fellow
> Department of Biological Sciences
> Macquarie University, New South Wales, Australia 2109

> Email: pwil...@bio.mq.edu.au
>
> >>> Mat Bond <matjb...@hotmail.co.uk> 09/01/09 3:17 AM >>>

[Mat Bond]

Michelle Greve.

Thanks,

I believe so; I am aiming primarily to separate out the subsoil and
topsoil salinity variables,

Cheers,

Mat

[Mat Bond]

Greetings,

I have created an access query with the fields MU_Global and S_ECE in
a .dbf format.

However, arc view GIS 3.2 will not display the data in the table when
I open it, instead it shows a blank table.

I have also tried to make the same table as you made (MU_Global,
SU_Symbol and SU_Code columns) with the same result, a blank table
being produced.

If I could just create a map of the world distribution of the variable
S_ECE it would be enough for my needs.

Thanks very much for the help,

[Mat Bond]

also I can enable the Spatial Analyst Toolbox in arc GIS 3.2 but
cannot find the button to show it

[michelle g]

Hi Mat,
have you tried opening the dbf table with Excel or text, to see if the
table looks ok? Make sure there are no spaces, numbers or funny
characters in the row names.
I haven't worked with ArcGIS 3.2, so I'm not sure if that may be a
problem.
Anyway, if your table is OK, but you cannot see in ArcGIS, you can
email it to me and I can see if I can open it here.

The Spatial Analyst tool should be displayed as one of the toolboxes.
(I don't know how well you know ArcGIS - look for a little red chest
on your toolbar to display the toolboxes, if you can't see them.) You
mus also make sure that your licence actually allows you to use
Spatial Analyst.
Michelle

[Mat Bond]

Yeah, the row names look fine (although some of the MU_GLobal values
don't have any associated S_ECE values)

I have been able to get hold of ArcGIS 9.3 now and the query table
works fine, however when i try and perform a link on it i get the
following message:

Executing: AddJoin hwsd.bil Value F:\HWSD_v110\Data\HWSD_RASTER
\SALINITY.DBF MU_GLOBAL KEEP_ALL hwsd.bil
Start Time: Fri Oct 09 16:19:25 2009
ERROR 999998: Unexpected Error.
Failed to execute (AddJoin).
End Time: Fri Oct 09 16:19:25 2009 (Elapsed Time: 0.00 seconds)

I think the problem may be that i am just opening the hwsd data file
without changing it to a grid format and i get a message telling me
that the file is missing data reference information when i opening in
ARc map, how would I go about exporting it to a grid format?

I am realy quite new to Arc GIS im afraid but i have found the toolbox
and the tools all seem to be present in the 9.3 version, thank you
very much for your help

Mat

[Mat Bond]

Yeah I think the problem is that im not converting to a grid format,
is there a specific utility for this?

[michelle g]

Hi Mat
1. It doesn't matter that you get the message 'the file is missing
data reference information' when you open in ArcGIS. All it means is
that the projection is not defined. ArcGIS still reads it fine and you
can analyse it as needed. (You can (and should) when getting your
final layer define the same projection of the original file. You can
do that using the 'Project' tool. The best way to do that is, when you
are have opened the tool and you choose the 'Output Coordinate
System', is by then selecting 'Import' and browisng until you find the
original soil file which should have the right projection. (But, as I
said, you might as well wait until you have created the final layer
before you do this.))
2. To export to grid, use the tool 'Raster to Different Format', and
select 'Grid' as the format.

I hope this helps you along.
Michelle

[Mat Bond]

Great, thanks very much I will it try again.

[Mat Bond]

Thanks for all the help, I have performed the link and am now getting
a message informing me that there is no spatial annalist tool-kit
licence, I think I will have to use another data source as it does not
seem that I will be able to get hold of a licence.

Thanks again

Mat

[Christian Wicaksono]

This helped me a lot. 

Michelle thanks a lot for your instructions. although maybe worth to mentioned:

One need to obtain the MS access database from http://www.iiasa.ac.at/Research/LUC/External-World-soil-database/HTML/HWSD_Data.html?sb=4, along with the zipped .bil, to be able to create the Query table. That's what got me confused in the beginning (or maybe I just didn't read your instructions properly). 

Anyway,  Thanks again 

Cheers, 

Christian 

[Natalia Quinteros]

Thanks for sharing your method! this post was very helpful!

Regards, 

Natalia Q.  

[Li Xu]

Hi, all

I have a question on HWSD. Please, give me an answer.

Will, HWSD is very good for studying, but it is hard to subset the ROI, you know.

I have installed the tool—HWSD Viewer, it could not subset.

Next, it was imported to ArcGIS 9.3, however, losing the geographical information.

So, how can I subset HWSD with my shapefile in ArcMap?

Any help could be appreciated.

Thanks a lot!

Merry Christmas Eve!

 

Cheers,

LI Xu

24 December, 2013

在 2009年9月1日星期二UTC+8上午1时16分01秒，Mat Bond写道：

[David Galbraith]

Where is your study area? Lots of data sets exist that can be used in Maxent, but it can be helpful to do a bit of web searching prior to processing your environmental descriptors for Maxent. Have you worked with that soil data much in the past?

--
You received this message because you are subscribed to the Google Groups "Maxent" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maxent+un...@googlegroups.com.
To post to this group, send email to max...@googlegroups.com.
Visit this group at http://groups.google.com/group/maxent.
For more options, visit https://groups.google.com/groups/opt_out.

[Li Xu]

Thanks a lot for your reply. Oh, it's my fault, it can be subset by ArcGIS.

Thanks!

2013/12/28 David Galbraith <davidmg...@gmail.com>

--
You received this message because you are subscribed to a topic in the Google Groups "Maxent" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/maxent/Hqe-FG5EX80/unsubscribe.
To unsubscribe from this group and all its topics, send an email to maxent+un...@googlegroups.com.