Land Use Classification and Deforestation Detection in Southwestern - PowerPoint PPT Presentation

Land Use Classification and Deforestation Detection in Southwestern Nicaragua B Y : C A S S A N D R A R I V A S , O P H E L I A W A N G , A N D B O S T E V E N S N O R T H E R N A R I Z O N A U N I V E R S I T Y

Introduction/Background  Rapid land use change in the tropics  Deforestation  Fragmentation  Environmental degradation  Determining current land use and temporal change is important  Sustainable conservation and restoration planning

Research Objectives and Questions  Objective: To delineate areas of vulnerability to change in SW Nicaragua, specifically areas of deforestation and natural regeneration between the years of 2000-2009 in order to inform future conservation and restoration efforts in the region.  What are the land use patterns revealed in the land use/cover classification?  How have the land use patterns changed since the year 2000?  What are the current vulnerable areas of change?

Methods: Study Area  Rivas Isthmus  Paso del Istmo  Ecosystems include: Dry-tropical forests, moist forested areas, and coastal mangroves  Other land cover types include: Pasture, plantation, crops, and urban (Map sources: ArcMap 10.1; Paso Pacifico.org)

Methods: Field Data Collection & Classification  Ground Referencing Data Imagery Collection SPOT, collected Jan-Feb 2009  Digitize roads and access Classification points Manual training polygons  398 data points collected; ENVI 4.7 RuleGen (Quest) assessed cover data at each Final Classes: point  Wetland  Urban  Crop  Pasture  Plantation  Young Regrowth  Old Regrowth  Young Secondary Forest  Old Secondary Forest Accuracies and Kappa statistics

Methods: Change Detection & Cluster Analysis  Land-use/cover classification 2000 (Sesnie et al. 2008)  Post-classification change detection in ENVI  Classification preparation for change detection analysis  Plantation  Hotspots/cold spots of change  Cluster and Outlier Analysis using Moran’s I  Neighborhood clustering score

Final 2009 Classification • Over all accuracy is 87.68% • Producers accuracies range 76.16-95.63% • Users accuracies range 76.34 - 95.52 % • Khat scores – overall and all individual classes except for wetland-mixed and old regrowth are in strong agreement

Forest Change (2000-2009) Forest Change 60 No Change 50 40 Change (%) 30 20 10 0 Crop Pasture Regrowth Plantation Forest Class Type

Forest to Pasture • Hotspots of change occur in the Northcentral part of the isthmus • Also in the southern region where last large tracks of forest occur despite rugged terrain • Cold spots occur throughout region

Forest to Regrowth • Hot spots occur in pockets along the rugged Pacific mountains • Cold spots occur in throughout the intense topography • At some point this forest was converted into another class and now regrowing

Forest to Plantation • Hotspots occur in a few pockets to the north and in the agricultural lake side area Cold spots occur in the southern • area where old secondary forest remain*

Regrowth Change (2000-2009) Regrowth Change 50 45 40 35 No Change (%) 30 Change 25 20 15 10 5 0 Crop Pasture Regrowth Plantation Forest Class Type

Regrowth to Pasture • Change is scattered throughout and is prolific • Reveals the strong influence of this class in the area • Hot spots are specifically located in the north central region

Regrowth to Forest • Hotspots are taking place in the southern part of the isthmus - Important for increasing forest near other older forest - Reveals the change in the area despite rugged topography

Discussion  Change Detection  50% of the secondary forest was converted into other classes  Deforestation is still occurring at a high rate approximately 5.6% per year  Regrowth is converted mainly to pasture (prevailing force)  However, is also maturing into secondary forest (23%)

Conclusions  Future conservation efforts should utilize this data for a more comprehensive conservation assessment  Deforestation and rapid land use change is continuing to occur throughout this dynamic landscape  Conservation and restoration efforts must be efficient given limitations  Cluster (hotspot/coldspot) analysis helped in defining specific areas of change  This study is a unique example of leveraging resources and collaboration to assess land use change for practical conservation

Thank you  NAU & the Lab of Landscape Ecology and Conservation Biology  My advisors Ophelia Wang and Tom Sisk  Paso Pacifico  My volunteers and Nicaraguan field guides