Advanced

Mapping Primary Production for the West African Sahel with Satellite Data

Seaquist, Jonathan LU (2001)
Abstract (Swedish)
Popular Abstract in Swedish

Kartering av primärproduktion i Västrafrikanska Sahel genom användning av satellitdata.



I avhandlingen har utvecklats en Light Use Efficency (LUE) modell som medger kartering av bruttoprimärproduktionen (GPP) exemplifierad i ett område i Västafrikanska Sahel. Modellen utnyttjar satellitbaserad månatlig Normalised Difference Vegetation Index (NDVI) tillsammans med andra typer av data. Bildkvaliteten hos dessa månadsdata (Maximum Value Composites, MVC) kan vara låg vilket visas genom användandet av en förbättrad geostatistisk metod för uppskattning av brus i bilddata. Vidare visas att datakvaliteten kan förbättras genom användandet av nya metoder för att skapa kompositbilder.... (More)
Popular Abstract in Swedish

Kartering av primärproduktion i Västrafrikanska Sahel genom användning av satellitdata.



I avhandlingen har utvecklats en Light Use Efficency (LUE) modell som medger kartering av bruttoprimärproduktionen (GPP) exemplifierad i ett område i Västafrikanska Sahel. Modellen utnyttjar satellitbaserad månatlig Normalised Difference Vegetation Index (NDVI) tillsammans med andra typer av data. Bildkvaliteten hos dessa månadsdata (Maximum Value Composites, MVC) kan vara låg vilket visas genom användandet av en förbättrad geostatistisk metod för uppskattning av brus i bilddata. Vidare visas att datakvaliteten kan förbättras genom användandet av nya metoder för att skapa kompositbilder. Dessa metoder använder de termiska kanalerna hos NOAA AVHRR samt svepvinkelinformation för att reducera molneffekter och inflytande från ogynnsamma svepvinklar. Dessa förbättrade data utnyttjas sedan tillsammans med annan information för att kartera säsongsmässig GPP genom användande av den utvecklade LUE modellen. I LUE modellen reduceras den komplexa biologiska processen till ett hanterbart antal variabler och parametrar. För att kompensera för bristfälliga stationsdata utnyttjas molninformation (CLAVR) hos NOAA AVHRR data för att uppskatta olika parametrar beskrivande markytans energibalans, bl. a. inkommande fotosyntetiskt aktiv strålning (PAR). Kvoten mellan inkommande och absorberad PAR beräknas ur NDVI och multipliceras med PAR för att ge mängden absorberad fotosyntetiskt aktiv strålning. En korrektionsfaktor för vegetationens fuktighetsstress beräknas ur en hydrologisk modell som behandlar transpiration och avdunstning från bar mark separat. Denna korrektionsfaktor används för att reducera den fotosyntetiska aktiviteten i LUE modellen. Den absoluta precisionen hos de uppskattade GPP-värdena minskar med vegetationens täthet medan den relativa precisionen ökar. Primärproduktion uppskattad genom LUE modellen är systematiskt högre för tät vegetation jämfört med oberoende modelleringar med modellen CENTURY men är jämförbar med uppskattningar hämtade ur litteraturen. Den beskrivna LUE modellen kan användas för att generara information relevant för en lång rad miljöproblem i torrområden. (Less)
Abstract
A light Use Efficiency (LUE) model is developed that allows the mapping of total growing season Gross Primary Production (GPP) for the West African Sahel, using the Normalized Difference Vegetation Index (NDVI) together with other data. Image quality may be poor in monthly NDVI maximum value composites as shown by an improved geostatistical noise estimation technique. Quality may be improved by other compositing methods that use NOAA AVHRR-derived surface temperature and scan angle information to reduce residual cloud amount and off-nadir bias. These data are then used in conjunction with ancillary information to map total growing season GPP using the LUE approach, which reduces the complexities of plant growth to a simple parametric... (More)
A light Use Efficiency (LUE) model is developed that allows the mapping of total growing season Gross Primary Production (GPP) for the West African Sahel, using the Normalized Difference Vegetation Index (NDVI) together with other data. Image quality may be poor in monthly NDVI maximum value composites as shown by an improved geostatistical noise estimation technique. Quality may be improved by other compositing methods that use NOAA AVHRR-derived surface temperature and scan angle information to reduce residual cloud amount and off-nadir bias. These data are then used in conjunction with ancillary information to map total growing season GPP using the LUE approach, which reduces the complexities of plant growth to a simple parametric statement. To overcome the lack of ground data, NOAA AVHRR-derived CLAVR (CLouds from AVHRR) fields are used to derive several key parameters of energy balance, including Photosynthetically Active Radiation (PAR). Fraction of absorbed Photosynthetically Active Radiation (FPAR) is calculated from the NDVI and multiplied with PAR to yield Absorbed Photosynthetically Active Radiation (APAR). A water stress scalar is estimated with a two-layer hydrological model that treats separately bare soil evaporation and transpiration. This scalar is used to reduce potential photosynthetic capacity in the LUE model, as defined by the product of APAR and the potential growth efficiency. The absolute precision of GPP estimates decreases for dense vegetation while the relative precision increases. LUE primary production estimates are systematically higher for dense vegetation when compared to point estimates from the CENTURY model. This bias is not apparent when compared to previous work reported in the literature. The LUE model may be used to address issues related to desertification, food security, and climate change. (Less)
Please use this url to cite or link to this publication:
author
opponent
  • Professor Eastman, Ron, Dept. of Geography, Clark University, Worcester, Mass., USA
organization
publishing date
type
Thesis
publication status
published
subject
keywords
Physical geography, Sahel, CENTURY, Monte Carlo, GPP, transpiration, LUE, CLAVR, PAR, compositing, MVC, geostatistics, noise, NDVI, NOAA AVHRR, geomorphology, pedology, cartography, climatology, Fysisk geografi, geomorfologi, marklära, kartografi, klimatologi
pages
198 pages
publisher
Department of Physical Geography, Lund University
defense location
Sölvegatan 13, 3rd floor lecture hall
defense date
2001-09-21 10:15
ISSN
0346-6787
ISBN
91-973857-2-7
language
English
LU publication?
yes
id
d64b0c6c-d8b6-479b-8256-689b828dce02 (old id 41870)
date added to LUP
2007-08-01 08:44:39
date last changed
2018-05-29 09:50:09
@phdthesis{d64b0c6c-d8b6-479b-8256-689b828dce02,
  abstract     = {A light Use Efficiency (LUE) model is developed that allows the mapping of total growing season Gross Primary Production (GPP) for the West African Sahel, using the Normalized Difference Vegetation Index (NDVI) together with other data. Image quality may be poor in monthly NDVI maximum value composites as shown by an improved geostatistical noise estimation technique. Quality may be improved by other compositing methods that use NOAA AVHRR-derived surface temperature and scan angle information to reduce residual cloud amount and off-nadir bias. These data are then used in conjunction with ancillary information to map total growing season GPP using the LUE approach, which reduces the complexities of plant growth to a simple parametric statement. To overcome the lack of ground data, NOAA AVHRR-derived CLAVR (CLouds from AVHRR) fields are used to derive several key parameters of energy balance, including Photosynthetically Active Radiation (PAR). Fraction of absorbed Photosynthetically Active Radiation (FPAR) is calculated from the NDVI and multiplied with PAR to yield Absorbed Photosynthetically Active Radiation (APAR). A water stress scalar is estimated with a two-layer hydrological model that treats separately bare soil evaporation and transpiration. This scalar is used to reduce potential photosynthetic capacity in the LUE model, as defined by the product of APAR and the potential growth efficiency. The absolute precision of GPP estimates decreases for dense vegetation while the relative precision increases. LUE primary production estimates are systematically higher for dense vegetation when compared to point estimates from the CENTURY model. This bias is not apparent when compared to previous work reported in the literature. The LUE model may be used to address issues related to desertification, food security, and climate change.},
  author       = {Seaquist, Jonathan},
  isbn         = {91-973857-2-7},
  issn         = {0346-6787},
  keyword      = {Physical geography,Sahel,CENTURY,Monte Carlo,GPP,transpiration,LUE,CLAVR,PAR,compositing,MVC,geostatistics,noise,NDVI,NOAA AVHRR,geomorphology,pedology,cartography,climatology,Fysisk geografi,geomorfologi,marklära,kartografi,klimatologi},
  language     = {eng},
  pages        = {198},
  publisher    = {Department of Physical Geography, Lund University},
  school       = {Lund University},
  title        = {Mapping Primary Production for the West African Sahel with Satellite Data},
  year         = {2001},
}