Discriminating disturbance from natural variation with LiDAR in semi-arid forests in the southwestern USA

T. L. Swetnam, A. M. Lynch, Donald Falk, Stephen Yool, Phillip Guertin

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Discriminating amongst spatial configurations and climax size of trees in forests along varying physical gradients from time since last disturbance is a significant component of applied forest management. Understanding what has led to the existing vegetation's structure has important implications for monitoring succession and eco-hydrological interactions within the critical zone: the near-surface environment where rock, soil, air, and biota interact and regulate ecosystem services. This research demonstrates the utilities of local indicators of spatial association (LISA) to (1) quantify natural variation in forest structure as derived from aerial Light Detection and Range (LiDAR) across topographically complex landscapes at ecologically relevant scale, i.e., individual trees; (2) map previously recorded but poorly defined forest disturbances; and (3) link scalable topographic indices to observed tree size distributions. We first selected a priori undisturbed and disturbed stands scanned by aerial LiDAR that included: preservation, mechanical fuels treatment, and logging in two similar semi-arid forest ecosystems. Next, we compared two topographic indices: the topographic position index (TPI) and topographic wetness index (TWI) to climax tree height and two related LISAs: the Getis-Ord G<inf>i</inf> and Anselin local Moran's I. G<inf>i</inf> and I measure local spatial clustering, producing z-scores that represent the significance of each statistic. The tallest trees in each study area were found to be located on negative TPI and greater TWI values (valley bottom positions and higher wetness conditions, respectively). Stands of trees with positive z-scores (i.e., overstory trees with similarly tall neighbors) were most likely to be located in the a priori undisturbed areas. Disturbed locations, on the other hand, were defined by neighborhoods of negative z-scores (i.e., shorter trees with short neighbors). Future applications include locating and reconstructing historical disturbances and developing a consistent strategy for characterizing stands for forest inventory monitoring.

Original languageEnglish (US)
Article number97
JournalEcosphere
Volume6
Issue number6
DOIs
StatePublished - Jun 1 2015

Fingerprint

disturbance
climax
monitoring
soil air
soil biota
overstory
forest inventory
vegetation structure
detection
forest trees
ecosystem services
forest ecosystems
logging
forest management
ecosystem service
forest ecosystem
statistics
valleys
index
rocks

Keywords

  • Anselin local Moran's I
  • Disturbance ecology
  • Forest structure
  • Getis-Ord G
  • LiDAR
  • Local indicators of spatial association
  • Topographic indices

ASJC Scopus subject areas

  • Ecology, Evolution, Behavior and Systematics
  • Ecology

Cite this

Discriminating disturbance from natural variation with LiDAR in semi-arid forests in the southwestern USA. / Swetnam, T. L.; Lynch, A. M.; Falk, Donald; Yool, Stephen; Guertin, Phillip.

In: Ecosphere, Vol. 6, No. 6, 97, 01.06.2015.

Research output: Contribution to journalArticle

@article{2e81263faf5548c485f73729746563ed,
title = "Discriminating disturbance from natural variation with LiDAR in semi-arid forests in the southwestern USA",
abstract = "Discriminating amongst spatial configurations and climax size of trees in forests along varying physical gradients from time since last disturbance is a significant component of applied forest management. Understanding what has led to the existing vegetation's structure has important implications for monitoring succession and eco-hydrological interactions within the critical zone: the near-surface environment where rock, soil, air, and biota interact and regulate ecosystem services. This research demonstrates the utilities of local indicators of spatial association (LISA) to (1) quantify natural variation in forest structure as derived from aerial Light Detection and Range (LiDAR) across topographically complex landscapes at ecologically relevant scale, i.e., individual trees; (2) map previously recorded but poorly defined forest disturbances; and (3) link scalable topographic indices to observed tree size distributions. We first selected a priori undisturbed and disturbed stands scanned by aerial LiDAR that included: preservation, mechanical fuels treatment, and logging in two similar semi-arid forest ecosystems. Next, we compared two topographic indices: the topographic position index (TPI) and topographic wetness index (TWI) to climax tree height and two related LISAs: the Getis-Ord Gi and Anselin local Moran's I. Gi and I measure local spatial clustering, producing z-scores that represent the significance of each statistic. The tallest trees in each study area were found to be located on negative TPI and greater TWI values (valley bottom positions and higher wetness conditions, respectively). Stands of trees with positive z-scores (i.e., overstory trees with similarly tall neighbors) were most likely to be located in the a priori undisturbed areas. Disturbed locations, on the other hand, were defined by neighborhoods of negative z-scores (i.e., shorter trees with short neighbors). Future applications include locating and reconstructing historical disturbances and developing a consistent strategy for characterizing stands for forest inventory monitoring.",
keywords = "Anselin local Moran's I, Disturbance ecology, Forest structure, Getis-Ord G, LiDAR, Local indicators of spatial association, Topographic indices",
author = "Swetnam, {T. L.} and Lynch, {A. M.} and Donald Falk and Stephen Yool and Phillip Guertin",
year = "2015",
month = "6",
day = "1",
doi = "10.1890/ES14-00384.1",
language = "English (US)",
volume = "6",
journal = "Ecosphere",
issn = "2150-8925",
publisher = "Ecological Society of America",
number = "6",

}

TY - JOUR

T1 - Discriminating disturbance from natural variation with LiDAR in semi-arid forests in the southwestern USA

AU - Swetnam, T. L.

AU - Lynch, A. M.

AU - Falk, Donald

AU - Yool, Stephen

AU - Guertin, Phillip

PY - 2015/6/1

Y1 - 2015/6/1

N2 - Discriminating amongst spatial configurations and climax size of trees in forests along varying physical gradients from time since last disturbance is a significant component of applied forest management. Understanding what has led to the existing vegetation's structure has important implications for monitoring succession and eco-hydrological interactions within the critical zone: the near-surface environment where rock, soil, air, and biota interact and regulate ecosystem services. This research demonstrates the utilities of local indicators of spatial association (LISA) to (1) quantify natural variation in forest structure as derived from aerial Light Detection and Range (LiDAR) across topographically complex landscapes at ecologically relevant scale, i.e., individual trees; (2) map previously recorded but poorly defined forest disturbances; and (3) link scalable topographic indices to observed tree size distributions. We first selected a priori undisturbed and disturbed stands scanned by aerial LiDAR that included: preservation, mechanical fuels treatment, and logging in two similar semi-arid forest ecosystems. Next, we compared two topographic indices: the topographic position index (TPI) and topographic wetness index (TWI) to climax tree height and two related LISAs: the Getis-Ord Gi and Anselin local Moran's I. Gi and I measure local spatial clustering, producing z-scores that represent the significance of each statistic. The tallest trees in each study area were found to be located on negative TPI and greater TWI values (valley bottom positions and higher wetness conditions, respectively). Stands of trees with positive z-scores (i.e., overstory trees with similarly tall neighbors) were most likely to be located in the a priori undisturbed areas. Disturbed locations, on the other hand, were defined by neighborhoods of negative z-scores (i.e., shorter trees with short neighbors). Future applications include locating and reconstructing historical disturbances and developing a consistent strategy for characterizing stands for forest inventory monitoring.

AB - Discriminating amongst spatial configurations and climax size of trees in forests along varying physical gradients from time since last disturbance is a significant component of applied forest management. Understanding what has led to the existing vegetation's structure has important implications for monitoring succession and eco-hydrological interactions within the critical zone: the near-surface environment where rock, soil, air, and biota interact and regulate ecosystem services. This research demonstrates the utilities of local indicators of spatial association (LISA) to (1) quantify natural variation in forest structure as derived from aerial Light Detection and Range (LiDAR) across topographically complex landscapes at ecologically relevant scale, i.e., individual trees; (2) map previously recorded but poorly defined forest disturbances; and (3) link scalable topographic indices to observed tree size distributions. We first selected a priori undisturbed and disturbed stands scanned by aerial LiDAR that included: preservation, mechanical fuels treatment, and logging in two similar semi-arid forest ecosystems. Next, we compared two topographic indices: the topographic position index (TPI) and topographic wetness index (TWI) to climax tree height and two related LISAs: the Getis-Ord Gi and Anselin local Moran's I. Gi and I measure local spatial clustering, producing z-scores that represent the significance of each statistic. The tallest trees in each study area were found to be located on negative TPI and greater TWI values (valley bottom positions and higher wetness conditions, respectively). Stands of trees with positive z-scores (i.e., overstory trees with similarly tall neighbors) were most likely to be located in the a priori undisturbed areas. Disturbed locations, on the other hand, were defined by neighborhoods of negative z-scores (i.e., shorter trees with short neighbors). Future applications include locating and reconstructing historical disturbances and developing a consistent strategy for characterizing stands for forest inventory monitoring.

KW - Anselin local Moran's I

KW - Disturbance ecology

KW - Forest structure

KW - Getis-Ord G

KW - LiDAR

KW - Local indicators of spatial association

KW - Topographic indices

UR - http://www.scopus.com/inward/record.url?scp=84939189768&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84939189768&partnerID=8YFLogxK

U2 - 10.1890/ES14-00384.1

DO - 10.1890/ES14-00384.1

M3 - Article

AN - SCOPUS:84939189768

VL - 6

JO - Ecosphere

JF - Ecosphere

SN - 2150-8925

IS - 6

M1 - 97

ER -