Create line features between point features within a given distance

[PatrickGreber]

Good morning everybody!

I'm currently working on a small GIS-Project regarding Network-Analysis. First I wanted to develop a network between certain wetland areas. I have created a point feature class in the centre of each wetland polygon. Now i try to connect these points with lines to create a network I can work with. Is there a tool that is able to connect each point feature with other point featues in a distance of 1300 Meters through a line?

In a next step I would like to adapt the distance regarding ecological and topographical information, through a cost-distance or multi-criteria analysis. Would this tool from step one be able to integrate such a cost-distance analysis and only connect point features within an adapted distance?

I'm really new to this with little expirience and grateful for any support =)! 

Thank you in advance for your help 😃

[DavidPike]

It's a bit dated, but the technique should still be the same

Connect point to closest feature in ArcGIS 10.2 - Geographic Information Systems Stack Exchange

generate a near table and then create the lines from that. (I guess you delete any table values over 1300 before you generate the lines)

As to the other questions, it might be better to state what you want to achieve rather than us helping you with a solution you've proposed.

[PatrickGreber]

Yes, that would totally help! As a part of a school exercise, I need to do a network analaysis and calculate centrality measures of a network. But first I have to generate such a network by myself.

My idea was to create a network of wetland habitats that are connected through lines. A habitat is only connectet if in reach of the dispersal distance of a certain butterfly (1300m in my case). In the end, I hoped to get a network with connected habitats and that I'm able to calculate centralities of each area. Furthtermore, I wanted to affect the dispersal distance by other features like streets, woodlands or lakes (travelling through these habitats cost more). The task with the centrality is given by my school and needs to be implemented.

So my ultimate goal would be a network of wetlands that are connected with each oter dependent on dispersal distance and topographical barriers.

Does this help=)?

Thank you for your time and help!

[DavidPike]

The double-post is a forum blip.  I was sure I replied initially, then it disappeared and I reposted.
You might want to check out the Cost-Distance toolset.  

It would take a little work to create a cost raster initially, but then you can create cost-distance rasters from each habitat, also run least-cost path analysis from each habitat, then possibly append the results to the edges (connecting lines) of your network.

Understanding cost distance analysis—ArcGIS Pro | Documentation

[PatrickGreber]

Yes! That's an idea I was already working on. Could it be possible to add a cost that represents the distance? As an Example: Create a buffer around each point and cells within each cell have no cost an cells outside the buffer have high cost values? Than run the least cost ath analysis?

Thank you for your time, I really appreciate it!

[DavidPike]

You'd be best-off producing  cost surface for your entire study area. This will be the 'input cost raster' for most of the other tools. 

You don't need to think about the distance in the cost raster, it calculates that from the source (the output is the sum of the multiples of distance travelled and cost-raster cell values).  Set cells of no impedance as 1, then impedant cells such as woodland or water, set them to 2 or 3 or even 50 etc.  (set NoData to introduce a barrier cell.) depending how you rank their relative impedance.

A cost-distance analysis would sum up the least-cost path distance from the source (habitat) to each output cell.  you could also just set a maximum cost-distance cutoff.  I'd recommend spending some quality time reading up on the toolset.

[PatrickGreber]

Hi David,

Sorry for my late response, I had some exams going on. Now I'm back at this problem. I generated a cost raster as you said (picture 1, I only included 3 different raster layers to work with until I know all about the workflow). Then I run the cost-distance analysis with it (picture 2), I used my habitat point layer as starting points. I'm still not quite sure if I'm able to use that many starting points or f I'm supposed to do such an anaysis for each individual point (which would be quite time consuming with that mny points). The problem I face now is, that my startig points are also my ending points. If I want to calculate the least cost path, the output will only be one cell for each point, because it returs immideatly..

And by doing this, I only get the output of one least cost path right? But for my future network, I would like to have all  paths from one starting point to ohter points within 1300m and including cost..

Sorry for the trouble, but this is all kind of new to me!

Thank you for your help=)

[DavidPike]

You could generate a near table, get rid of any values over 1300 and then use the XY to line tool to generate the edges you need.

What's your end-goal with this?