shows 
MULTITEMPORAL REMOTE SENSING FOR ASSESSMENT OF LAKE BIOPRODUCTION IN RELATION TO 
INFLUENCING FACTORS 
Kerstin Munier, Sigrid Roessner, Sabine Thiemann; GeoForschungsZentrum Potsdam, Germany 
Commission VII, Working Group No. 2 
KEY WORDS: Satellite remote sensing, IRS-1C, multitemporal, GIS, lake water, landuse 
ABSTRACT 
In this paper water bodies of the Northern Brandenburg lake district, situated about 80 km northwest of Berlin, are 
investigated by means of satellite remote sensing based on field measurements of optical lake water properties. These 
properties reflect the intensity of bioproduction of algae which significantly influences the trophic state of lakes. 
Multitemporal remote sensing allows the assessment of typical seasonal behavior of bioproduction in lakes during the 
observed period from August 1996 until September 1997. Additionally, the multitemporal IRS-1C satellite data are 
analyzed for factors influencing the lake water properties, such as lake size and surrounding landuse. These information 
are combined with non-remote sensing data using a GIS for further analysis of possible relationships between lake 
water properties and influencing factors. 
KURZFASSUNG 
In dieser Arbeit werden Seen des Nordbrandenburgischen Seengebietes, das sich etwa 80 km nordwestlich von Berlin 
befindet, mit Methoden der Satellitenfernerkundung unter Nutzung von Feldmessungen zu ausgewählten 
Gewasserparametern untersucht. Die optischen Eigenschaften der Seen reflektieren deren Bioproduktion, die den 
Trophiezustand wesentlich bestimmt. Die multitemporale Fernerkundung erlaubt die Erfassung typischer saisonaler 
Veránderungen im Gewásserzustand wáhrend des Beobachtungszeitraumes zwischen August 1996 und September 
1997. Zusátzlich werden die multitemporalen IRS-1C Daten in Bezug auf Faktoren, die die Wassereigenschaften 
  
beeinflussen, wie z. B. Seegröße und umgebende Landnutzung, ausgewertet. Diese Informationen werden mit 
nichtfernerkundlichen Informationen in einem GIS kombiniert, um sie auf mögliche Beziehungen zwischen 
Wassereigenschaften und beeinflussenden Faktoren auszuwerten. 
1. INTRODUCTION 
The Northern Brandenburg lake district covers an area 
of about 4000 km” and contains more than 1000 lakes 
of different trophic states. Local environmental 
authorities monitor present conditions. The assessment 
of water quality of these lakes requires time consuming 
field measurements. Remote sensing has the potential 
to support this process in a spatially continuous way. 
Existing remote sensing based studies (e.g., Baban, 
1993; Mayo et al, 1995) analyze the relationship 
between spectral properties of lake waters and 
parameters for trophic state estimation like chlorophyll- 
a content. 
The visible (VIS) and near infrared (NIR) part of the 
spectrum of different multispectral satellite-based 
(Landsat-TM, SPOT, ADEOS) and hyperspectral 
airborne sensors (CASI, AVIRIS) have been used for 
the assessment of water quality. In case of satellite 
systems only spectral information from two to three 
broad bands have been available for these 
investigations. However, for Landsat TM data good 
correlation between chlorophyll-a concentration and 
the index (TM1-TM3)/TM2 was found by Mayo et al. 
(1995) for lake Kinneret in Israel. The sensitivity of 
differentiation is not optimal because of the broad 
bandwidth. 
In the area of the Northern Brandenburg lake district 
hyperspectral investigations using a field spectrometer 
(FieldSpec FR, ASD Inc.) and data of the 
hyperspectral airborne imaging spectrometer CASI 
(tres Inc.) have been carried out during the same 
period of time (Thiel et al., 1997; Thiemann et al., 
1998). They show strong relationships between field 
spectrometer and imaging spectrometer data on the 
one side and field measurements of secchi disk 
transparency and chlorophyll-a content on the other 
side. These results allow a quantitative estimation of 
both parameters from spectral information. 
In this study data of the multispectral Indian Remote 
Sensing Satellite IRS-1C LISS 3 (tab.1 ) are analyzed 
for their potential for multitemporal monitoring of the 
water quality during one year of bioproduction between 
August 1996 and the end of September 1997 (tab. 2). 
The acquired multitemporal dataset of IRS-1C data 
containing 6 scenes is analyzed to determine the 
relationship between spectral reflectance of the three 
VIS/NIR bands of the IRS-1C instrument and the lake 
parameters chlorophyll-a content and secchi disk 
transparency. In Northern Brandenburg the influence of 
suspended matter on the spectral signal can be 
neglected because of the low relief and the related low 
current of the water. Satellite remote sensing intends to 
show the spatial variation of optical lake water 
properties within the study area and the seasonal 
development of bioproduction in the lakes during the 
observed period. It offers the opportunity 
simultaneously map the situation of many lakes in an' 
efficient way. 
The trophic state of lakes is influenced by a complex 
combination of parameters (llmavirta et al, 1989, 
Hóntzsch et al., 1996). The influence of the individual 
parameters on the lake properties varies between the 
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 129