Using Remotely Sensed Data and the FAREAST Forest Succession Model - - PowerPoint PPT Presentation
Wildlife Conservation Society Using Remotely Sensed Data and the FAREAST Forest Succession Model to Estimate Biomass and Leaf Area Index (LAI) Across a Complex Landscape in the Russian Far East CITES-2009 NEESPI Workshop Krasnoyarsk, Russia
Wildlife Conservation Society
CITES-2009 NEESPI Workshop Krasnoyarsk, Russia July 15, 2009
Tiger range, 1900 and today (current or potential)
Varied topography, climate and vegetation
A vast landscape shaped by disturbance. Tigers and leopards generally avoid conifer forests, burned areas and human activity (Miquelle et al 2005).
Seasonal fires in agricultural areas
Highways and railroads cross wildlife areas. Logging, both small scale and commercial.
Anthropogenic development
+ 331-393 adults and sub- adults (20 – 36 mos.) in the Russian Far East (2005 census (Miquelle et al 2007)). + No more than 15 – 20 Amur tigers in China (Chinese gov’t 2007, Yang et al 1998, Sun et al 1998) + Endangered species (IUCN/World Conservation Union 2008) + Numbers are stable to slightly decreasing (Miquelle et al 2005) + Greatest threats – loss of habitat, insufficient prey, poaching (primarily for bones and fur).
Amur leopard (left) and Amur tiger (right) photographed by infrared camera trap in Primorskiy Krai, RFE. Photo – Wildlife Conservation Society The principle prey of Amur tiger and Amur leopard are ungulates – red deer (Cervus elaphus), sika deer (C. nippon), wild boar (Sus scrofa), and roe deer (Capreolus capreolus pygargus).
+ Primary prey of the Amur tiger (Miquelle et
al,1999)
+ Similar to the elk (or wapiti) in North America + Most often found in riverine areas (Korean pine, oak, birch, other deciduous vegetation) + Avoid spruce/fir and larch forests + Eat stems, twigs, leaves of broadleaved trees and shrubs, herbs and sedges, forbs, lichens, fruits, fungi. Also eat willows, poplar, mountain ash, oak, cowberries, and blackberries
+Tigers range wherever wild pigs are found
(Miquelle et al 1999)
+ Pig numbers may be declining because of poaching (Stephens et al 2006) + Mostly found in Korean pine forests + Eat Korean pine nuts, plant roots, acorns, soil invertebrates, carrion + Range is limited by deep snow
+ May be expanding range northward and displacing red deer as climate warms (Stephens et al 2006) + Range limited by deep snow + Approx 117 kg (male) & 73 kg (female), i.e. 1/2 - 2/3 the size of red deer + Found in oak forests. + Eat bark, twigs, buds, acorns. In summer, eat fungi, herbs
pine needles, later juniper, algae (for salt).
and in clearings (Heptner et al, 1988).
C. Capreolus pygargus by Komarov in A.N. Heptner 1988
+ Ecologically very adaptable to different habitats. + Mostly range north of tiger habitat + In RFE, presence is associated with fir and larch forests, which tigers avoid
Ungulate (hooved animals) presence is associated with Korean pine (Pinus korainsis), and broadleaf temperate forests, especially those including oak (Quercus mongolica). (Miquelle et al 1999) Pine nuts and acorns are important food sources for ungulates and small mammals, especially during harsh winters.
Alexander Omelko, PhD, Institute of Biology and Soil Science, Far Eastern Branch of the Russian Academy of Sciences,
Prey Species – Vegetation Relationship
Vegetation type Red deer Wild boar Sika deer Roe deer Moose Riverine (oak-birch, Korean pine-deciduous, spruce-fir)
Fir
Summary of relationships between prey presence and forest type, based on track encounter rate. “+” = tracks encountered most frequently “ – “ = habitats avoided (Blank space means forest type is used in accordance with its availability (Stephens et al 2006)).
landscape representing Amur tiger and Amur leopard habitat.
based interpretations of forest types, forest structure (LAI), biomass and canopy height.
are best suited for climate change simulation in the study area.
assessing predator-prey-vegetation relationships using resource selection function analysis.
disturbance based on current forest composition and characteristics.
single location and demonstrates forest succession leading to mature tree stands.
– Characteristics and requirements related to growth, mortality and regeneration for 44 tree species – Site characteristics, such as elevation, soil moisture and nutrients – Climate parameters, such as temperature and precipitation
across an elevational gradient at Changbai Mountain in northeastern China and simulated forest composition and successional patterns in terms of biomass at 23 of 31 sites in Russia.
versus observed NPP at 593 Forest Survey Stations in China.
Days
200 plots of 0.05 ha 1 site For each year at one site or point, 200 plots of 0.05 hectares (500 m2) are run. Temperature and precipitation vary randomly within constraints of average monthly standard deviation for these characteristics.
Actual Versus Observed Basal Area by Species at Four Elevations y = 0.8546x R2 = 0.8539
5 10 15 20 25 30 35 5 10 15 20 25 30 35
Actual Data Model Prediction
Tests of the FAREAST Model on Changbai Mountain gradients
1 to 1 line
5 10 15 20 25 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Temperature (C)
Observed GFCM21 MPEH5 NCCCSM HADCM3
End of 20th century modeled and observed monthly temperature (C) at Primorskiy Krai weather stations. (T. V. Loboda, University of MD, Dept. of Geography, 2009)
20 40 60 80 100 120 140 160 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Precipitation (mm) Observed GFCM21 MPEH5 NCCCSM HADCM3
End of 20th century modeled and observed precipitation (mm) (T. V. Loboda, University of MD, Dept. of Geography, 2009)
Legend
Size of pie slices indicates the biomass composition of mature forests across the NEESPI region.
By running the FAREAST model (200 simulated plots for 700 years starting with an open plot) for 234 weather stations in the NEESPI region, one
(MODIS land cover product: IGBP
2007))
Fig 1. Two InSAR images developed from Dual-pol L-band Synthetic Aperture Radar (SAR) data for the same field study location show the rough terrain in the region, which causes difficulties in digital image
generated from the InSAR data, are being used to make terrain corrections and perform geo-coding of the radar data. (G. Sun, Univ.
Land Use (ILU) data
(bands 4,3,2) Landsat Thematic Mapper image
at 988 points representing Geoscience Laser Altimeter System (GLAS) data locations.
precipitation for each point were extrapolated using National Climate Data Center database.
using allometric equations and then predicted from GLAS waveform indices using neural network. (PI –
' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
Arhar Norsk Ternej Czekunda Poronajsk Habarovsk Vladivostok Mys_Zolotoj Pogranichnyi Blagovescensk Dal'nerechensk Im.Poliny_Osipenko Ekaterino-Nikol'skoe Aleksandrovsk-Sahalinskij
We also ran the FAREAST model at 1,000 random points. At each point, the model was calibrated for temperature, precipitation, elevation and soil characteristics. For each of these points, Leaf Area Index (LAI) calculated by FAREAST was compared with a MODIS- derived estimate (Myneni et al., 2002.)
Height vs Biomass
50 100 150 200 250 300 5 10 15 20 25 30 35 40 Height Biomass TCha-1 Fareast Sun GLAS
cover.
closure.
1,000 Random Points - LAI Difference (Remotely sensed - FAREAST simulation)
Negative Y values suggest forest disturbance, such as logging or agriculture, since Fareast simulates an old growth forest.
0.00 1.00 2.00 3.00 RP Dif LAI UMD - FE
FAREAST model output – mature forest Remotely sensed data – actual land cover For example: Remotely sensed LAI of 1 (very little forest) – FAREAST-derived LAI of 6 (full forest) ___________________________________
Field work was conducted from September 18, 2007 – October 22, 2007, to collect a data set of diameter (dbh), tree height, basal area and sapling growth statistics for trees and saplings in 10 30-meter circles at 12 sites, for a total of 120 study locations.
Lena Pimenova, Chief Scientist and Associate Director, Sikhote-Alin Biosphere Reserve/Zapovednik (SAZ) and Mikhail Gromyko, PhD, Sr. Scientist, SAZ, measuring tree height in Sikhote-Alin Zapovednik (SAZ).
biomass, leaf area index (LAI) and forest structure.
a large, inaccessible landscape.
characteristics, and site and environmental factors to simulate mature stands.
detected by comparing remote sensing results with modeled results.
sensing.
Anatoly Astafiev – Director, Sikhote-Alin Zapovednik/Biosphere Reserve Sveta Bonduchuk - Sikhote-Alin Zapovednik/Biosphere Reserve Mikhail Gromyko - Sikhote-Alin Zapovednik/Biosphere Reserve Tatiana Loboda - University of Maryland, Dept. of Geography, College Park, MD Yelena Pimenova – Sikhote-Alin Zapovednik/Biosphere Reserve Dale Miquelle - Wildlife Conservation Society, New York; Russia Program, Terney, Russia Guoqing Sun - University of Maryland, Dept. of Geography, College Park, MD Tim Stevens - University of Virginia, Dept. of Environmental Sciences, Charlottesville, VA Ivan Csiszar - NOAA/NESDIS Center for Satellite Applications and Research, Camp Springs, MD