Instrumentation and remote sensing

 

Scientific investigators:

Prof. Dr. Thomas Nauss

Falk Hänsel

Stephan Wöllauer

Spaska Forteva (assoc.)

(Uni Marburg)

Prof. Dr. Christoph Kleinn

Dr. Paul Magdon

Dr. Stefan Erasmi (assoc.)

Collins Kukunda (assoc.)

Wanda Graf (assoc.)

(Uni Göttingen)

Martin Fellendorf

(Uni Ulm)

Matthias Gross

(TU München)

Frank Suschke

(SGN Frankfurt)

Introduction

The core project „Instrumentation and remote sensing“ is responsible for both (i) the provision of systematic and large-scale measurements and recordings of meteorological and pedological parameters in all exploratories and (ii) the provision of remote sensing based, area wide information on the land cover and land use. Similar to the existing instrumentation infrastructure, a central archive of remote sensing products will be set up.

 

Description of the key activities related to the station network

Meteorological and pedological parameters are among the main factors of organismic diversity and functionality of ecosystem cycles. A systematic and large-scale measurement and recording of these environmental parameters is performed in all exploratories with more than 300 measurement stations.

Figure 1. Climate-station type AEMU at Hainich (Photo: F. Hänsel)
Figure 2. Tower at Swabian Alb (Photo: F. Hänsel)
Figure 3. Workshop for validation - at SEG43 (Photo: F. Hänsel)

Climate stations

At each of the three biodiversity exploratories, 100 climate stations have been installed on grassland research plots and inside the forests. Regarding the huge amount of stations, not all stations have been equipped in an equal manner but a gradual setup has been used. As of today, the following stations are operational:

  • 279 Core Environmental Monitoring Units (CEMUs))
  • 21 Enhanced CEMUs (EEMUs)

 

CEMU – this is the basic setup which is installed on all experimental plots on grassland and forest locations. Each CEMU measures air temperature and humidity (+200 cm), soil surface temperature (+10 cm), soil moisture (-10 cm, at VIP’s  additional -20 cm) and soil temperature (-5 cm, -10 cm, -20 cm, -50 cm) in intervals between 30 minutes and 1 hour.


EEMU - on 21 selected locations, the basic CEMU setup is extended by four component radiation balance, photosynthetic active radiation, wind speed and direction as well as precipitation sensor with a recording interval of 10 minutes. Some locations are also recording barometric pressure.
A visualization of the air temperature recordings can be found: hier

Observation towers

At the biosphere reserves Schorfheide-Chorin and Swabian Alb, two observation towers with a height of 37 m and 44 m respectively have been installed. Air temperature and humidity profiles, as well photosynthetic active radiation profiles are recorded in addition to precipitation and the above canopy radiation budget and wind dynamics.
2 more towers, installed by the Max-Planck-Institute for Biogeochemistry, can be used in the Hainich national park.

Data handling and processing

The station network records about 45.000 values per day within each exploratory. The data stream is transmitted to Marburg for the pre-processing and quality checks and uploaded to the BExIS data base afterwards. Because of the increasing amount of data, a new climate database component has been developed which will be implemented into BExIS soon.

 

Description of the key activities related to Remote Sensing

Background

In reaction to the increasing demand for remote sensing products and services from different research groups, the core project “Instrumentation” was extended by a remote sensing component in the current project phase and is now named “Instrumentation and Remote Sensing”. With the extended core project, a better coordination of the remote sensing activities within the Biodiversity Exploratories is envisaged. The remote sensing sub-project will provide area wide information on the land cover and land use of the three Biodiversity Exploratories which will complete the meteorological data collected by the instrumentation project. By providing and updating different maps derived from remote sensing data we want to enable also ‘non-remote sensing experts’ to integrate spatial and temporal components into their analysis to gain knowledge on functional biodiversity relationships at different scales.

Sensors

To cope with the manifold research questions and methods of the Biodiversity Exploratories, remote sensing data from a wide range of sensors will be acquired and pre-processed. This includes high resolution optical satellite data (e.g. Pléiades, RapidEye) as well as airborne hyperspectral and LiDAR sensors as shown in Table 1.

Table 1. Overview of the remote sensing products that will be used in the 4th phase of the Biodiversity Exploratories (2014-2017)

 

System

Plattform

Spatial resolution

# Bands

Spectral

resolution

Temporal resolution

Extent

RapidEye

Satellite

5m

5

0.44 – 0.85

2009-2015

~min. three phenological periods per year

area-wide

Landsat*

Satellite

15m – 120m

4 - 8

0.43 – 12.5

1972 - 2014,

~ yearly

area-wide

MODIS**

Satellite

250m – 1km

36

0.40 – 14.39

2001 to 2014,

~ daily

area-wide

Pléiades

Satellite

50cm – 2m

5

0.43 – 9.40

Once (2015)

3x100 km2

Hyperspectral

Aircraft

1m

>200

0.40 – 2.40

Once (2015)

3x100 km2

LiDAR

Aircraft

~14 Points/m2

Full wave form

---

Once (2015)

3x100 km2

*The Landsat data set includes MSS, TM, ETM+ und OLI/TIRS sensors.

**MODIS data from Terra and Aqua Satellites

 

Activities

Neighborhood analysis (land use and land cover mapping)
Ecosystem processes influencing the biodiversity are present at very different scales. It is hypothesized that several of the processes described for the intensively monitored EP’s and VIP’s are also influenced by the plot surroundings. In this work package, maps that describe the current land cover and land use of the Biodiversity Exploratories will be created. By that we wish to link the intensive observations of the EP’s and VIP’s to the larger landscape context.
The land cover classification key will be adapted to the ecosystem type. In forests, various tree species groups will be differentiated. For grass lands, agricultural land uses and shrubs will be the guiding factors for classification.

Figure 4. False color composite forest areas in the Biodiversity Exploraorie Hainich acquired with AISA Eagle/Hawk Sensors at July 2012 (Source: personal communication Mr Henning Aberle, Chair of Forest Inventory and Remote Sensing, Göttingen, ForestHype project DLR).

Landscape structure analysis
The maps from work package I allow quantitative analyses of the horizontal landscape structure of the three Biodiversity Exploratories. Linking it with the height information obtained from the LiDAR and Pléiades data analyses, we will extend the 2D information on horizontal landscape configuration by a vertical component to provide a more complete 3D description of landscape and vegetation structures.

Time series analysis
The relationships between ecosystem functions and land use intensities are among the main research topics in the Biodiversity Exploratories program. Often, the currently observed status is the result of processes that have been present over long time periods. Thus, the historical development of the landscape may contribute explaining the currently observed status. Remote sensing offers the possibility to make such descriptions using time series analysis techniques – when corresponding imagery is available.

 

By opening the Landsat archive to the public, NASA is providing a data store that can be used to build time series starting from the 1980ies. The objective of this work package is to acquire, pre-process and analyze Landsat images to construct temporal trajectories which describe the history of land cover changes in the Biodiversity Exploratories. In a second step we will analyze the potential of these change maps to be part of the explanatory features that serve as proxy for land use intensity.

Figure 5. Left: spatial distribution of the forest areas in the Biodiversity Exploratory Schorfheide. Right: Proportion of the different patch sizes (ha) of the total forest area
Figure 6. Workflow of a time series analysis of Landsat Satellite images for mapping changes and disturbances

To top