Spatial and temporal Spatial and temporal changes in Namaqualand Namaqualand: : changes in Separating out climate from Separating out climate from human- -induced changes induced changes human Ravi Kum ar Kum ar Vadapalli Vadapalli, Richard Knight & , Richard Knight & Ravi Eugene Moll Eugene Moll
INTRODUCTION & BACKGROUND INTRODUCTION & BACKGROUND Vegetation change using remote- -sensing is sensing is Vegetation change using remote based on based on � Reflectance ( � Reflectance ( Greeness Greeness) ) < < Unhealthy Healthy > > < < Unhealthy Healthy > > < < Low Cover NDVI High Cover > > High Cover > > < < Low Cover NDVI � Canopy Fragmentation � Canopy Fragmentation
Degradation in Succulent Karoo Karoo??? ??? Degradation in Succulent � Replacement Replacement � of palatable of palatable with with unpalatable unpalatable spp. . spp � Alteration in Alteration in � Seasonal Seasonal Change Change Tim e Tim e Series Series X Change in X Change in X Fragmentation in X Fragmentation in amount of Green amount of Green Canopy Cover Canopy Cover Vegetation Vegetation
METHODS METHODS � SKEP planning SKEP planning � domain domain Springbok Springbok � Time Series Time Series � Analysis (April Analysis (April 1998- - Feb 2003) Feb 2003) 1998 Garies Garies tri- -monthly monthly tri � SPOT VGT SPOT VGT � Imagery 1 km x 1 Imagery 1 km x 1 Vanrhynsdorp Vanrhynsdorp km NDVI km NDVI � Standardized PCA Standardized PCA �
METHODS METHODS SPOT VGT Data SPOT VGT Data � April 1998 April 1998- - Feb 2003 Feb 2003 � (after that the sensor (after that the sensor was re- -calibrated) calibrated) was re � Imaging integrated for 10 Imaging integrated for 10 � day periods – – 177 177 day periods images images � NDVI calculated NDVI calculated �
Principle Components Principle Components � Reduce our 177 images into Reduce our 177 images into � fewer independent “ “views views” ” fewer independent � Analogous to identifying what Analogous to identifying what � objects are by slicing various objects are by slicing various sections, e.g. What fruit is this? sections, e.g. What fruit is this? Anom alies Anom alies C3 C1 C2 Geographical Geographical Seasonal Seasonal Banana Banana
Components Components 1 - - 78.7% 78.7% 1 2 - - 6.7% 6.7% 2 3 - - 2.6% 2.6% 3 � Component Loadings Component Loadings � (correlation between (correlation between original S10 data and original S10 data and each component) each component) plotted plotted � Component Graphs Component Graphs � compared to expected compared to expected graphs graphs
RESULTS – – Rainfall Rainfall RESULTS 140 GARIES KAMIESKROON 120 SPRINGBOK STEINKOPF 100 Rainfall (mm) 80 60 40 20 0 JAN 1998 JAN 1999 JAN 2000 JAN 2001 JAN 2002 JAN 2003 Year
RESULTS – – Component Loading 1 Component Loading 1 RESULTS 1.2 1 Apr'02 Jan'03 Apr'99 Apr'01 Component Loadings Apr'98 0.8 7 8 .7 % 0.6 Apr'00 0.4 Extended 0.2 Mining Mining Dry Period 0 Geographical Geographical Features Features Agriculture Agriculture 98% explained 98% explained
RESULTS – – Component Loading 2 Component Loading 2 RESULTS 1 0.8 Apr'00 Sum m er Sum m er 0.6 C om ponent Loadings Rainfall Rainfall 0.4 W inter W inter Apr'01 Rainfall Rainfall 0.2 Apr'99 Apr'02 Apr'98 0 Jan'03 -0.2 -0.4 -0.6 Seasonal Seasonal Perm anent Perm anent Effects Effects Agriculture Agriculture
RESULTS – – Component Loading 3 Component Loading 3 RESULTS 0.3 0.2 Jan'03 Apr'99 0.1 Apr'98 Com ponent Loadings Apr'01 Apr'02 0 -0.1 Apr'00 -0.2 -0.3 -0.4 Mining Mining -0.5 -0.6 Major Climate Major Climate Anomalies – – Rainfall / Rainfall / Anomalies Agriculture Agriculture Drought events Drought events
DISCUSSION DISCUSSION Vegetation is stress-sensitive • Component 1 • Component 1 • Graph: Show two “ “crashes crashes” ” in in Greeness Greeness values values • Graph: Show two • Map: Identified Mining & Permanent Agric. impacts • Map: Identified Mining & Permanent Agric. impacts • Component 2 • Component 2 Climate Regimes Discernable • Graph: Annual cycles of Greeness Greeness • Graph: Annual cycles of • Map: Winter/Summer rainfall divide • Map: Winter/Summer rainfall divide • Component 3 Component 3 • Droughts identifiable Graph: One extreme “ “Crash Crash” ” in Greeness in Greeness Graph: One extreme Map: Areas “ “independent independent” ” of Greeness of Greeness “ “crash crash” ” Map: Areas
ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS MAPOSDA Project (EU) MAPOSDA Project (EU) VITO for SPOT imagery VITO for SPOT imagery
Recommend
More recommend
