Tree inventory system

From Wildsong
Jump to navigationJump to search

This doc very much under construction; in fact not even really a doc yet. More of a stream of consciousness.

Brian Wilson 12:48, 23 February 2006 (PST)

What is a tree inventory?

A 100% tree inventory is a tree-by-tree portrait of the urban forest. The inventory is the foundation for planning all work in the urban forest. This information provides a tool for both day-to-day and long term management.

Why use a tree inventory system?

  • Budget and report tool
  • Work planning and asset management
  • Hazard tree removal
  • Tree historical info
  • Pruning cycle
  • Planting program
  • Enforcement of vegetation code

Eugene uses its system to track compliance when a developer is required to plant trees.

Eugene tracks trees now from the time they are installed by coding each tree with a unique id tag. This allows trend tracking; for example, if a group of trees all have a similar problem, they can see if they all came from one supplier.

What data do we collect?

For each tree,

  1. Location (address)
  2. Species
  3. Diameter / Age
  4. Condition (dead, poor, fair, good, excellent)
  5. Structural Defects (topped, included bark, decay cavities.. Etc)
  6. Planting site (park strip width, tree well size)
  7. Proximity to utilities (0verhead power lines, sidewalk condition) ETC.

Goals

  1. Create a complete tree inventory system.
  2. Trick the City of Corvallis and OSU into beta testing it. :-)

I am evaluating tree inventory systems because (1) Corvallis needs one (2) the project requirements are interesting to me.

It has to have a field data collection component (for example, ArcPad based) with or without GPS. In my mind, with GPS.

It has to have a database backed component to manage the data, and that database contains spatial data so integration with GIS is critical.

It should have an asset management component, interesting to me because it is like the software release tracking system I was writing a few years ago.

In short, this project pulls together just about everything I am interested in these days: GPS, GIS, field work, and spatial databases.

Potential partners

Three obvious ones, they are my neighbors: City of Corvallis, Oregon State, City of Albany. So far, Corvallis and OSU have expressed interest. Have not contacted Albany yet.

Requirements

I will develop a better sense of what is required by evaluating the software and by some meetings with the staff at City of Corvallis and at Oregon State. Both entities have expressed interest and both are local to me.

Field requirements

Data collection by interns/contractors. Needs to be fast and simple.

Access to all data in field for work crews and arborist.

Management requirements

Basically this is an asset management system integrated with GIS. It has to be able to do the usual GIS stuff plus tracking plus reporting.

Although this is possibly the most important component I am relegating it to the back burner for now.

Build or buy?

Here are a few notes on tree inventory systems that are available.

Or I should say were available in 1997; this doc is being updated now. It is based on material pulled from web sites and things have changed since it was written. When developing feature requirements though, even documentation on products no longer available can be useful.

My current inclination is to put together a system by combining some commercial and some free software to create a useful data collection system. I would posit that one requirement for any purchased commercial system would be the ability to work with existing data. So by embarking on data collection before installing a complete asset management system, we're not losing anything.

Client/server or Web?

At a glance most of these systems are client server based. Client server is ideal for commercial software, you have to buy and install and maintain a licensed copy of the client on each workstation.

I really dislike client-server, because you have to buy and install and maintain software on multiple machines! :-)

For the tree inventory system I think a hybrid approach is most appropriate. I want to leverage the strength of using ESRI ArcPad and ArcGIS Desktop, but I want a web-based component that will allow the workflow to be unimpeded by the requirements of having many copies of expensive software installed all over town whether they are needed or not.

I want someone doing fieldwork to be able to download the latest files into ArcPad without having to learn how to operate ArcGIS Desktop. I don't want a GIS tech tied up building ArcPad projects every time data collection work is needed.

I want public access to (portions of) the system so that anyone can look up information on tree projects; for example, a citizen could get all available information via web when one of those scary signs appears attached to the tree in front of his house.

I want anyone at a front desk setting anywhere in the city to be able to pull up the complete unabridged data when someone walks in the front door asking for information.

Implementation

Field component

Data collection field component based on a customized ArcPad on a PDA.

Advanced version either a portable ArcMap or ArcPad on a laptop.

Optional Bluetooth GPS

Treeworks uses ArcPad and ArcGIS Desktop. Eugene uses ArcPad on a couple Dell Axims and on a Trimble GeoXT to do field data collection. They use maps and air photos in ArcPad more than GPS to enter the location of trees when entering data.

My idea for GPS is to use it to help you bring up the right extent on the screen as a convenience, not to determine tree location. You'd leave the GPS on until you reach the site, so that the correct view is on your screen when your feet hit the ground.

By its nature, you will ALWAYS be under a tree when using this system. :-) Then again there are those dandy $20000 laser range finders...

Server component

Mapserver (or ArcIMS) based application

We have ArcIMS available, but I lean towartds the opensource mapserver since basing the server component on free stuff brings down the final price by 50% should it ever get used anywhere else.

If I build it on a mapserver base I can test it out on top of shapefiles and then transition to PostGIS system, which should give much better performance for queries.

I want a user to be able to browse to an area of interest on a web-based map interface, and then use the currently viewed map to define the extent for queries, reports, and export to ArcPad.

I've looked at ArcPad project files, and this really seems pretty doable.

This means the checkout of data could be done without a copy of ArcGIS being involved.

Locking... can I lock an extent in PostGIS? That would be very cool. Someone could check out an extent to ArcPad, then upload new data into the system. What about revision control?

At least for the tree database, these operations should be fairly easy as long as the trees are stored as points. Only the tree inventory will be updatable from ArcPad; other layers would be clipped for export but not locked.

Redlining would be nice. Just the ability to write notes on the map and keep that in a separate layer. I suppose storing photos the same way would be nice.

Desktop component

The desktop app is just ArcGIS. Use ArcMap to maintain spatial data.

Links

Information on iTree software: http://urbanforestrysouth.org/Resources/Library/TTResource.2005-01-11.0622/view

International Society of Arborists, who do the "Trees are good" website

Society of American Foresters publishes Journal of Forestry [June 2005] has an article on CMT's PC-GIS

Corvallis Micro Technology PC-GIS ($500)
QGIS is a "user-friendly Open Source Geographic Information System" (free)
Thuban is "an interactive geographic data viewer". (free)