Difference between revisions of "LANDIS"

Revision as of 21:06, 19 September 2010

General System description

System Name: LANDIS

Brief overview

LANDIS is a spatially explicit landscape simulation model. It models natural processes, such as fire, wind, and insect disturbance; succession, and seed dispersal, as well as forest management.

The purpose of LANDIS is to predict forest landscape change over long time periods. LANDIS was developed because previous forest simulation models were not suited for larger areas. Spatially explicit computer simulation of entire landscapes requires simulation at appropriate spatial resolution and level of mechanisms and processes that are modelled.

LANDIS is a spatially explicit landscape simulation model. It models natural processes, such as fire, wind, and insect disturbance; succession, and seed dispersal, as well as forest management. LANDIS is raster-based and tracks only presence or absence of tree species in a given age class. Disturbance events are simulated stochastically based on mean return intervals and disturbance size. [1]

System origin

LANDIS development began in the early 1990s under the direction of Dr. David Mladenoff, with partial funding from the Northern Research Station, Rhinelander, Wisconsin. Dr. Hong He joined the team in the late 1990s followed by Dr. Robert Scheller several years later. There are two currently supported versions of LANDIS which were developed with primary funding from the Northern Research Station of the USDA Forest Service.

LANDIS 4.0 is a fully modular software product with improved fire simulation and new capabilities for simulating fuel accumulation and decomposition and disturbance by biological agents such as insects and disease. Dr. He led the development of this version at the University of Missouri in collaboration with the Northern Research Station in Rhinelander, Wisconsin.

LANDIS-II is a completely re-engineered version developed at the Forest Landscape Ecology Laboratory, University of Wisconsin-Madison, in collaboration with the Northern Research Station in Rhinelander, Wisconsin. LANDIS-II was designed to advance forest landscape simulation modeling in many respects, including:

allows for the incorporation of ecosystem processes and states (e.g., live biomass accumulation) at broad spatial scales
has flexible time steps for every process,
uses an advanced architecture that will significantly increase collaborative potential. ^[1]

Support for specific issues

Landscape planning

Support for specific thematic areas of a problem type

Silvicultural

Capability to support decision making phases

[Intelligence, Design, Choice, Monitoring ^[2]]

Intelligence: LANDIS supportS understanding natural changes in ecosystem states
Design: LANDIS allows creating various management simulations to understand how these affect ecosystems over time

Related systems

LANDIS-II has taken an open source, modular design for which numerous extensions are available at http://www.landis-ii.org/exts
Companion tools:^[3]
- LANDISVIEW - A graphical tool for quickly examining and animating LANDIS-II outputs
- PnET for LANDIS-II - A tool for calculating maximum ANPP and/or species probability of establishment.
- APACK Tool for Landscape Metrics and Analysis
- IAN: Extensible Image Analysis Tool (Calculates Landscape Metrics):

Data and data models

Typical spatial extent of application

Typically applied to large landscapes from thousands to millions of hectares.

Forest data input

LANDIS represents landscapes as a grid of cells and tracks age cohorts of each species (presence/absence or biomass) rather than individual trees^[4]. Spatial resolution is flexible, generally 10-500 m pixel sizes^[5].

Required LANDIS-II data can be broken into three basic data types: species data, ecoregion data, and initialization data.^[6]

Species data include the species life history attributes and species establishment probabilities. Life history attributes can be found in the literature and in the North American Silvic manual. They can also be determined from expert knowledge of the landscape of interest. Species establishment probabilities - the probability of survival after successful germination - are needed for each species and ecoregion combination. These can be estimated from expert knowledge, inventory data (how frequent is a species on each ecoregion?), or from multi-modeling (using a separate model to calculate species establishment probabilities).

Ecoregion definitions are highly dependent upon the landscape and the research or management question. In the midwest, ecoregions are often defined very broadly because of the low topography. Ecoregions are often defined using available landtype definitions: USDA Forest Service sub-sections, HUC, STATSGO polygons, etc. The extent of ecoregions also depends on the process of interest. If fine-scale fire behavior is being simulated, fine-scale ecoregions will be desirable.

Initialization data are the species and ages found on your landscape at the first time step. These data are often the most difficult and expensive to create. And again, the difficulty of creating these data will be dependent upon the data available. In the U.S., Forest Inventory and Analysis data is often summarized by ecoregion and cross-tabulated with satellite classification data. In some areas, more detailed stand cruise data is available for initialization. In general, the fewer species and the more even-aged the forest, the easier initialization will be. Boreal forests are easier, southern hardwoods more difficult.

Type of information input from user (via GUI)

Biophysical

Models

Forest models

Silvicultural
Climate impacts
Insects/Disease
Fire

Social models

Social and economic impacts must be inferred from spatially explicit estimates of biophysical change.

Decision Support

Definition of management interventions

[Define what is available for the manager to intervene in the forest: time of harvest, plantations, thinnings, reconversions...Existence of prescription writer, simple enumeration of all possibilities, scenario simulation , etc.]

Typical temporal scale of application

LANDIS is typically run for simulations from 10-1000 years.
LANDIS-II has flexible time steps for every ecological process.

Types of decisions supported

Decision-making processes and models

Simulation

Output

Types of outputs

Spatial analysis capabilities

Abilities to address interdisciplinary, multi-scaled, and political issues

System

System requirements

Operating systems:

Development status: Prototype being applied

Architecture and major DSS components

[Describe the basic architecture of the system in software and hardware. Desktop client-server, web based, as well as the integration with available systems. Basic data flow, focusing on retrieval of required input and propagation and implementations of decisions. Mention its modular and scalability capabilities.]

Usage

[Describe the level of use: Research level use, Industry use, Government use.]

Computational limitations

[Describe the system limitations: e.g. number of management units, number of vehicles, time horizon.]

User interface

User interface quality:

Complexity of system / user interface: Using this program requires a substantial investment in learning the details of the software.

Documentation and support

Support: A bulletin board for LANDIS users is available at: http://ash.forest.wisc.edu/cgi-bin/users/YaBB.pl

Installation

Prerequisite knowledge needed: This software is not designed for mass consumption. Successful application is probably best accomplished through a team that includes individuals with working knowledge of UNIX systems, C and C++ language, GIS packages, forest vegetation dynamics, associated resource interactions, and forest planning. [Mowrer]

Cost: Not priced

Demo:

References

Cited references

External resources

Website: http://www.landis-ii.org/

Publications: http://www.landis-ii.org/documentation/PublicationsPage

[1] ttp://nrs.fs.fed.us/tools/landis/

[2] ttp://fp0804.emu.ee/wiki/index.php/Simon%27s_decision_making_model

[3] ttp://www.landis-ii.org/users

[4] ttp://nrs.fs.fed.us/tools/landis/

[5] ttp://web.missouri.edu/~umcsnrlandis/

[6] ttp://www.landis-ii.org/documentation/Gentle%20Introduction/WhatDataWillINeed

[1]

[2]

[3]

[4]

[5]

[6]

@@ Line 64: / Line 64: @@
 LANDIS represents landscapes as a grid of cells and tracks age cohorts of each species (presence/absence or biomass) rather than individual trees<ref>http://nrs.fs.fed.us/tools/landis/</ref>. Spatial resolution is flexible, generally 10-500 m pixel sizes<ref>http://web.missouri.edu/~umcsnrlandis/</ref>.
+Required LANDIS-II data can be broken into three basic data types: species data, ecoregion data, and initialization data.<ref>http://www.landis-ii.org/documentation/Gentle%20Introduction/WhatDataWillINeed</ref>
+Species data include the species life history attributes and species establishment probabilities. Life history attributes can be found in the literature and in the North American Silvic manual. They can also be determined from expert knowledge of the landscape of interest. Species establishment probabilities - the probability of survival after successful germination - are needed for each species and ecoregion combination. These can be estimated from expert knowledge, inventory data (how frequent is a species on each ecoregion?), or from multi-modeling (using a separate model to calculate species establishment probabilities).
+Ecoregion definitions are highly dependent upon the landscape and the research or management question. In the midwest, ecoregions are often defined very broadly because of the low topography. Ecoregions are often defined using available landtype definitions: USDA Forest Service sub-sections, HUC, STATSGO polygons, etc. The extent of ecoregions also depends on the process of interest. If fine-scale fire behavior is being simulated, fine-scale ecoregions will be desirable.
+Initialization data are the species and ages found on your landscape at the first time step. These data are often the most difficult and expensive to create. And again, the difficulty of creating these data will be dependent upon the data available. In the U.S., Forest Inventory and Analysis data is often summarized by ecoregion and cross-tabulated with satellite classification data. In some areas, more detailed stand cruise data is available for initialization. In general, the fewer species and the more even-aged the forest, the easier initialization will be. Boreal forests are easier, southern hardwoods more difficult.
 === Type of information input from user (via GUI) ===