Dyna clue

[ah3881]

Does anyone know how to use Dyna clue to predict future land cover.
I seem unable to even input data, let alone run anything
Please could someone help?
Thanks so much!
Alice

[Angela Cuervo]

Dear Alice.

I don´t know how to use de Dyna, but I recomend you to check the Land Change Modeller of Idrisi Taiga (manual).... I use it to predict lang cover change. It uses Markov chains process, but it you can include variables potencially associated with the change.

Sincerely, 

--
Angela P. Cuervo-Robayo
Lic. en Biología, Universidad Distrital FJC
Est. Maestría en Ciencias/Medio ambiente
Universidad Autónoma de Baja California
México

[AliceUK]

Thanks

I do have Idrisi, but have not used it for 2 years (and have only used
it to prepare rasters for ENFA-which I used to use, and have forgotton
what little I did know!)
How do I go about predicting land cover change, in Idrisi

Thanks so much

Alice

[Florencia Sangermano]

Hi Alice,

The land change modeler is in Idrisi version 15 (Andes) and version 16 (Taiga) and also you can get it as an extension for ARGIS. If you have any of these, check the Tutorial as it has a step-by-step guide on how to do it. Also, for each panel you have an icon (it is a question mark at the top right) that you can click there for more info.
You can get a free trial at http://www.clarklabs.org/index.cfm that comes with the manual and tutorials ... it is a full version that expires in 15 days... but it will give you a sense of what you can do with it.

Flor

Thanks

Thanks so much

Alice

--
You received this message because you are subscribed to the Google Groups "Maxent" group.
To post to this group, send email to max...@googlegroups.com.
To unsubscribe from this group, send email to maxent+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/maxent?hl=en.

[ah3881]

OK I now have Idrisi Taiga-but only for 15 days....
So far I have only found bits that analyse differences between pre-
existing land-cover maps, and nothing to look at climatic change
(which I have).
I need to predict land-cover change based on human activity and
climate, and the "help" section on Idrisi has not enlightened me.
Please tell me how to model change-as I really need to get it done-and
with very little time, and right now I am just rather confused...
Thanks so much
Alice

On Feb 17, 1:56 pm, Florencia Sangermano <fsangerm...@clarku.edu>
wrote:

> Hi Alice,
>
> The land change modeler is in Idrisi version 15 (Andes) and version 16 (Taiga) and also you can get it as an extension for ARGIS. If you have any of these, check the Tutorial as it has a step-by-step guide on how to do it. Also, for each panel you have an icon (it is a question mark at the top right) that you can click there for more info.

> You can get a free trial athttp://www.clarklabs.org/index.cfmthat comes with the manual and tutorials ... it is a full version that expires in 15 days... but it will give you a sense of what you can do with it.

> For more options, visit this group athttp://groups.google.com/group/maxent?hl=en.- Hide quoted text -
>
> - Show quoted text -

[Florencia Sangermano]

Hi Alice,

I probably misunderstood your question. I thought you wanted to do land cover change projections...
The land change modeler (LCM) lets you evaluate current changes in land cover and project the changes into the future. The projection is based on an empirical determination of the relationship between current change (extracted from two land cover maps that you have to give as input) and a set of driver variables (variables that cause the change in your study area). That relationship can be determined in LCM trough Neural Networks or Logistic Regression. For each transition you want to project, LCM will give you a transition potential surface based on the results of NN or LR -- these are then combined to generate a soft prediction of change (vulnerability map) and then a hard prediction (scenario) can be created based on assigning the amount of land that will change in the future (based on the Markov transitions) to the areas of highest vulnerability.
In my opinion any model of land change (CLUE, Dinamica, LCM etc) are a valid for projections over a short time period, since the anthropogenic variables (that are the main drivers of change) can vary a lot in that time, and their relationship with the transitions can be affected. I would recommend you to go through the LCM tutorial to understand better how it works.

If what you want to do is to find the relationship between a particular land cover and climate conditions, I would do it in the neural network interface (called MLP) that is outside LCM...
You can run the model with current environmental variables to find the relationship, save the weights and then apply the weights to your future climate conditions.
Also why not doing that inside MaxEnt?, considering your land cover as a 'species' to be modeled.... that would be better than MLP since it is less of a black box, easier to run since you don't have to understand the complex parameters in the neural network...

Hope that helps-

Florencia

[ah3881]

Hello

I do, I am beginning to get the hang of the program-but how on earth
do I put in the legend files in the change analysis. I have actual
land cover cover in 1970, 1980 and 2000 which I could use to calibrate-
but I can't seem to input the legends-or find any instruction saying
how to go about doing that...
Thanks so much

Alice

ps. Doing it in Maxent may be difficult-as the land-cover would not be
in csv format

On Feb 18, 2:31 pm, Florencia Sangermano <fsangerm...@clarku.edu>

> > You can get a free trial athttp://www.clarklabs.org/index.cfmthatcomes with the manual and tutorials ... it is a full version that expires in 15 days... but it will give you a sense of what you can do with it.

> > For more options, visit this group athttp://groups.google.com/group/maxent?hl=en.-Hide quoted text -

[Florencia Sangermano]

See exercise 1-8 from tutorial-- In the Idrisi explorer when you click on an image you will see the metadata... scroll down until you see "categories" click there and add your legend.

Note that you land cover maps as should have equal number of categories, they should be numbered in sequence (with no empty categories) and should have exactly the same resolution and extent.

[ah3881]

I have been fighting the program all morning, don't you find it very
user unfreindly?
Even following the tutorials for the lcm all I get are messages such
as "failed, check integrity or memory" and if I attempt to proceed I
later get the message "catastrophic failure".
I input the 100 categorical variables 4 times, as Idrisi kept freezing
before I could save-and having to be closed in task manager .
When I had the relevant variables formatted, and with their legends
then all I got was messages, and not any results-I am completely
stumped-I have done everything according to its specifications but
just get errors...............................

On Feb 18, 3:32 pm, Florencia Sangermano <fsangerm...@clarku.edu>

> > > You can get a free trial athttp://www.clarklabs.org/index.cfmthatcomeswith the manual and tutorials ... it is a full version that expires in 15 days... but it will give you a sense of what you can do with it.

> > > For more options, visit this group athttp://groups.google.com/group/maxent?hl=en.-Hidequoted text -

[Florencia Sangermano]

Hi Alice,

I find it quite friendly :)... more so than other land change methods out there.
I'm sorry to hear that you are having difficulties with it. Since LCM is integrated inside the Idrisi GIS-RS software, the user should have some knowledge of this before diving into LCM... that is why the tutorial starts from displaying images and slowly advances up to LCM or time series analysis, more complex operations.

I use LCM all the time for REDD planning and to see the impact of deforestation of biodiversity and haven't experienced these problems, so I would guess it is a misunderstanding of the methodology and/or a hardware problem more than a software problem.

I would recommend that you contact Clark Labs' tech support about this.