INTERFACING GIS WITH A PROCESS BASED AGRO-ECOSYSTEM MODEL
–CASE STUDY NORTH CHINA PLAIN
Georg Baretha.* ,Zhenrong Yub
a
University of Cologne, Cologne, Germany,
b
China Agricultural University, Beijing, China
ABSTRACT:
The Sino-German Project between the China Agricultural University and the University of Hohenheim, Germany, focused on
sustainable Agriculture in the North China Plain. One major focus of the project was the establishment of an experiment field near
Beijing to investigate different agricultural practices and their impact on yield and environment. The second task was to set-up a GIS
based Agricultural Environmental Information System (AEIS) for the North China Plain (NCP), which exceeds almost the size of
Germany. Researchers from several departments are involved in the project: Agricultural Economics, Agricultural Informatics,
Vegetable Science, Landscape Ecology, Phytomedicine, Plant Nutrition, Plant Production and Soil Science. The major aim of the
AEIS for the NCP is to provide information (i) about agriculture in the region, (ii) about the impact of agricultural practices on the
environment and (iii) of simulation scenarios for sustainable strategies. Consequently, the AEIS for the NCP provides information for
decision support and therefore could be regarded as a Decision Support System (DSS), too. In this contribution, the focus is on the
importance of the linkage of process-based agro-ecosystem models, here the DNDC model with GIS. The purpose of the linkage is
the modeling of agro-environmental impacts on a regional level. Due to the geographic extent of the North China Plain, it could be
regarded as national level as well. A key issue in the GIS-based regional modeling is the establishment of an adequate geodatabase
which here is defined as a Agricultural Environmental Information System (AEIS). By using the database of the AEIS, it is possible
in this case study to model with a process-based model the emission of N2O, the volatilization of NH3, and the leaching of NO3- from
winter wheat/summer maize area for the entire North China Plain. Finally, the comparison with and evaluation of available models
and methods (IPCC; SLUSA) for regional calculation is presented.
1. INTRODUCTION
The Sino-German Project between the China Agricultural
University and the University of Hohenheim, Germany started
in November 1998 for 4½ years and was located in Beijing.
The focus of the project was on sustainable agriculture in the
North
China
Plain
(NCP)
(http://www.unihohenheim.de/chinaproject). Sustainable agriculture is a big
issue in China. One major focus of the project was the
establishment of an experiment field near Beijing to investigate
different agricultural practices and their impact on yield and
environment. Investigated crops and vegetables were winter
wheat, summer maize, spinach and cauliflower. For the crops,
three factors were considered in the different treatments. There
were three irrigation treatments (sub-optimal, conventional,
optimal), three N-fertilizer treatments (sub-optimal,
conventional, optimal), and two treatments of crop residues
(with and without residues). For the vegetables, two irrigation
treatments (conventional, optimal) and two N-fertilizer
treatments (conventional, optimal) were considered. Intensive
measurement equipment was installed on the experiment field
to collect long term data of soil water, volatilization, and
greenhouse gases. Plant and root data were sampled as well to
provide all necessary data for intensive modeling evaluation
and calibration. Additional investigations about pesticide,
herbicide and fungicide applications have been carried out on
the experiment field and in the outer Beijing area.
Questionnaires were used to collect information about
agricultural practices and the economical situation of the
farmers. Finally, plant and animal monitoring and mapping
were done to investigate the biodiversity in the study region.
Apart from the field experiment, the second task was to set-up
an Agricultural Environmental Information System (AEIS) for
the NCP (Bareth and Yu 2002), which exceeds almost the size
of Germany. In the sense of Bill (1999), an Environmental
Information System (EIS) is an extended GIS for the
description of the state of the environment referring to critical
impacts and loads. An EIS serves for the capture, storage,
analysis and presentation of spatial, temporal and attribute data
and provides basics for measures for environmental protection.
The establishment of an AEIS for the simulation of sustainable
scenarios means the set up of an extensive geo and attribute
database. Especially the modeling of the C- and N-cycles in
agro-ecosystems requires numerous input parameters like pH,
soil texture, fertilizer N-Input, animal waste input, use of
irrigation water, dates of sowing and harvest, yield, etc..
According to Bareth and Yu (2002), Fig.2 describes the
elements of an AEIS. Due to the definition of an EIS and the
defined aims of the AEIS for the NCP, an AEIS for sustainable
agriculture includes five different information systems which
are:
•
Base Geo Data Information System (BGDIS)
•
Soil Information System (SIS)
•
Climate Information System (CIS)
•
Land Use Information System (LUIS)
•
Agricultural
Management
Information
System (AMIS)
Most important for the spatial matching of all data in the AEIS
is the integration of an BGDIS. The BGDIS should provide
topographical data, elevation lines or a Digital Elevation Model
(DEM) and an administrative boundary data set. The SIS is
essential for providing soil parameters for the agro-ecosystem
modeling. Therefore, the SIS has to include (i) spatial soil
information in form of maps and (ii) a detailed description of
the soil types including soil genesis, physical and chemical soil
properties. The CIS provides the necessary climate/weather
information. Climate/weather maps can be generated from
point data using GIS interpolation methods. Land use data
should be provided by a LUIS. Detailed land use information
on crops should be stored in the LUIS. For detailed agroecosystem modeling, the information level in available land use
maps is rather poor. The analysis of multi-spectral,
hyperspectral and/or radar data from satellite or airborne sensor
is a standard method to retrieve such kind of information.
Finally, the AMIS is a crucial part. For the agro-ecosystem
modelling, farm management data like fertilizer N-Input,
animal waste input, use of irrigation water, dates of sowing and
harvest, yield, etc. are a must. The AEIS is the sum of the
described information systems. They are linked to each other
using GIS technologies. Additionally, models and methods for
data analysis have to be integrated in the AEIS. It is possible to
link or even integrate complex agro-ecosystem models into GIS
(Hartkamp et al., 1999) and consequently into the AEIS.
Consequently, the AEIS for the NCP provides information for
decision support and therefore could be regarded as a Decision
Support System (DSS), too. The major objectives of the AEIS
for the NCP are to provide information (i) about agriculture in
the region, (ii) about the impact of agricultural practices on the
environment and (iii) of simulation scenarios for sustainable
strategies.
Provinces of the North China Plain
Hebei
Tianjin
Shandong
Henan
Jiangsu
Anhui
100 200 300 Kilometers
Figure 1. Provinces and geographical boundary of the NCP
2. INTEGRATION OF AGRO-ECOSYSTEM MODELS
INTO GIS
Traditionally, process based agro-ecosystem models or
agronomic models in general are developed and used for site or
field scale. A major problem for regional applications of such
kinds of models is the limited regional data availability. The
scientific and technical progress in spatial sciences, especially
in GIS and remote sensing (RS) technologies, have enabled the
set up of soil, land use, climate and agricultural management
information systems. Unfortunately, complete data sets for
regions are hardly available and therefore the regional
application of agro-ecosystem models is still difficult. Lack of
data availability can nowadays solved by using again GIS and
RS technologies. Therefore, interfacing GIS with agroecosystem models is becoming more important.
Seppelt (2003) describes the coupling of GÌS and models in a
theoretical and technical context. Hartkamp et al. (1999)
introduce four different ways to interface GIS with agronomic
models. They also introduce definitions for the different ways
of interfacing. This is an essential point because the use of not
defined terminology causes confusion. Hartkamp et al. (1999)
define four terms which are frequently used as follows:
Interface:
Hartkamp et al. (1999) give a very good overview of examples
of interfacing GIS and models. They describe 46 publications.
While the interface type of 22 of the GIS-model interfaces is
linking, 23 represent combining and only 6 are of the
integrating type. None of the described integrating GIS-model
interfaces are for agro-ecosystem models focusing on C- and
N- dynamics. Most of them deal with topography, soil erosion
and groundwater flow.
N
Beijing
Anhui
Beijing
Hebei
Henan
Jiangsu
Shandong
Tianjin
0
The three different methods of interfacing GIS with models are
described as followed. Linking of GIS with models is basically
just an exchange of files or data. In this case, the model is
independent from the GIS and vice versa. Only the results of
each system are exchanged. Combining of GIS with models
involves processing of data and automatically exchange of data.
Finally, integrating GIS and models describes the real
incorporation of one system into the other.
The place at which diverse (independent)
systems meet and act on or communicate
with each other.
Link:
To connect.
Combine:
To unite, to merge.
Integrate: To unite, to combine, or incorporate into a larger
unit; to end segregation.
Especially for the regional modeling of C-and N-dynamics in
(agro-)ecosystems on a regional scale, the integration of such
models is important and can be regarded as a key issue.
Available approaches of GIS-model interfaces for agroecosystem models hardly use the GIS analysis capabilities. For
example, the Denitrification and Decomposition Model (DNDC)
(Li et al., 2001) has, in its latest versions, a regional model part
included. Spatial parameters like land use and soil are not
considered in their spatial relation. GIS is just used for result
display on county level using a county map. Plant (1998)
describes a GIS-DNDC-interface of the linking type for the
Atlantic zone of Costa Rica. This study shows clearly the
importance of interfacing agro-ecosystem models with GIS.
This introduced approach is not automated and consequently
very difficult to use for other regions. Therefore, we introduce
a method for integrating agro-ecosystem models into GIS
which enables fully automated spatial model simulations.
In this approach, we use the DNDC model for the integration
into a GIS. The DNDC Model was developed at the University
of New Hampshire in the early 90s (Li, 2000). Its original
purpose was the simulation of N2O and NO gas fluxes from
agricultural fields on a daily basis, but it also has a very
detailed soil C sub-model. It is currently the most used
simulation tool for N related gas fluxes and has been used for
county based national inventories of N2O for China (Li et al.,
2001), the USA (Li et al., 1996), the UK (Brown et al., 2002),
Germany (Werner 2003) and several more countries. The soil
organic matter sub-model is made up of three different soil C
pools with each being divided into a labile and resistant
fraction. The plant growth sub-model rather simply models
biomass at a potential growth rate which is modified by
environmental factors. The driving key variables are climate,
soil properties, and agricultural management. Even though it
has a very simple plant growth sub-model it is capable of
simulating trends in soil C quite accurately (Li et al., 1997). In
another simulation study of an arid farmland ecosystem in
China the DNDC model was also able to model the influence of
different management practices on soil C content (Liang and
Xu, 2000).
3. SLUSA-BASED INTERFACING OF GIS WITH
MODELS
The method for the integration of agro-ecosystem models is
based on the soil-land-use-system approach (SLUSA) (Bareth
et al., 2001). The SLUSA is based on the ecosystem approach
described by Matson and Vitousek (1990). This ecosystem
approach was furthermore disaggregated in order to estimate
and visualize the greenhouse gas emissions (CH4, CO2, N2O)
from agricultural soils for a distinct region (Bareth 2000). The
disaggregation of the ecosystem approach was undertaken by
using a GIS and available digital spatial data (Bareth 2000).
The SLUSA is a GIS and knowledge based approach for
environmental modeling. The first step of the SLUSA is the set
up of a GIS which contains relevant data. In the second step,
GIS tools are used to overlay climate, soil, land use,
topography, farm management, and other data like preserved
areas or biotope (if available). This procedure is the basis for
the spatial related identification of different soil-land-usesystems. In the third step, emission data and process knowledge
of the region of interest as well as from literature are linked to
these systems. For the linkage, N2O emissions potential classes
are created and knowledge based production rules are
programmed. The latter ones are commonly used to generate
new knowledge from expertise in knowledge based systems
like expert systems (Wright et al., 1993). Available data from
the region of interest are considered with higher priority.
Finally, GIS software tools are used in the fourth step to
visualize and to quantify the generated emission data. The
quantitative emission estimation is done by multiplying the
areas of each N2O emission potential class with the
representative value of each class. The new spatial emission
data can be integrated into other regional models like regional
farm models (Bareth and Angenendt, 2003).
Based on the SLUSA, Fig.2 describes the integration of agroecosystem models into GIS. While the steps 1, 2 and 4 remain
in the method, the third step of the SLUSA is changed. The
implementation of the knowledge based production rules of the
SLUSA is replaced by the integration of the agro-ecosystem
model. All input parameters for the model derive from the GIS
database. The advantage of this method is the automatic spatial
modeling. The GIS integrated DNDC model uses the input
parameters from the fields of the GIS database. Therefore, the
DNDC model program code was changed for that purpose and
necessary interfaces were programmed (Huber et al., 2002). For
the input, fixed field names were created in the DNDC model
and the GIS database must use these definitions. The results of
the model runs are automatically written into predefined fields
of the GIS database. Consequently, the results of the modeling
can immediately be used for regional calculation and
visualization by using GIS software tools. Finally, a new menu
button is integrated in ArcGIS which allows the user the easy to
use access for DNDC modeling.
Evaluation of models is a very important step. For this
procedure, measurement data is necessary. For Dongbeiwang
Township, the evaluation of the DNDC model can be done by
using measurement data from the experiment field. For
example, measurements of gas fluxes with the automatic
closed-chambers-technique are available from the research
project. The chambers and the automatic measurement system
are mainly designed by Martin Kogge of sub-project B2 of the
Sino-German
Project
(http://www.uni-hohenheim.de/
chinaproject/martin.htm). The chambers are designed to
measure CO2, CH4 and N2O. For the three different treatments,
measurements are done with three repetitions per plot. The
treatments are conventional fertilization and irrigation,
optimized fertilization and irrigation, and conventional
fertilization and optimized irrigation. N2O measurements are
available for summer maize 2001. For the traditional treatments,
the total emission for the period of 06-21-01 to 10-09-01 for
summer maize is 2.35 (± 0.48) kg N2O-N ha-1.
4. RESULTS
In Fig.3, the data of the AEIS for the North China Plain is
displayed. The official digital topographic database (1:250,000)
was purchased as Base Geodata Informationsystem (BGDIS).
On national level, the Institute of Geomatics in Beijing is
responsible for topographical data. In 1999, they finished the
digital topographical database on the scale 1:250,000 (DTD250)
in vector format. They provide the data as ArcInfo coverages
and many other GIS formats. The data set includes separate
files for rivers, lakes, roads, railways, cities, one degree net, 50
m contour lines and administrative boundaries (counties,
districts, provinces) (Bareth and Yu, 2004). The layers were
digitized from the topographical maps on the same scale. The
dates of the maps range from the early to the late eighties. A
DEM, generated from the contour lines, is available, too.
Web-GIS
User Interface
AEIS GDB
INPUTS
• Digital Topographical
Database 1:250.000
• Administrative Boundaries
• Contour Lines
• Landtype Map 1:500.000
• Some Soil Maps
1:200.000
• Expert Knowledge
AEISMaps
Maps
AEIS
Models
Base Geo Data Information System (BGDIS)
Basis GDB
BasisMaps
Maps
Basis
Soil Information System (SIS)
Soil GDB
SoilMaps
Maps
Soil
Climate Information System (CIS)
• Climate Data
Climate GDB
SLUSA Based Modeling of the C- and N-Dynamics in Agro-Ecosystems
2
1
1.
Land-Use-ID
Land-Use-ID
11
22
Land Use
il-ID
SSooil-ID
11
22
2
1
Soil
Geometric Overlay
2.
4
2
1
Tabular
Soil-Land-Use-System
-ID
LLaannddUUssee- ID
11
11
22
22
LLaannddUUssee
rasssslalanndd
GGra
rasssslalanndd
GGra
rabbleleLLaanndd
AAra
rabbleleLLaanndd
AAra
Integration of the DNDC
3.
Result-ID
Land Use-ID
Land Use
Soil-ID
Soil Type
1
2
3
4
1
1
2
2
Grassland
Grassland
Arable Land
Arable Land
1
2
2
2
Cambisol
Luvisol
Luvisol
Cambisol
4.
Visual
4
2
Overlay
Presentation
3
Emission Map
SSooil-il-ID
ID
11
22
22
11
Land Use GDB
•
•
•
•
Agric. County Statistics
Agric. Mgmt. Survey
Project Database
Literature
LandUse
UseMaps
Maps
Land
Agricultural Management Information System (AMIS)
Agric. Mgmt. GDB
Agric.Mgmt.
Mgmt.Maps
Maps
Agric.
SSooililTTyyppe e
CCaammbbisisool l
LLuuvvisisool l
LLuuvvisisool l
CCaammbbisisool l
Model into GIS
N-fertilization
-1 -1
in kg ha a
130
150
240
260
ClimateMaps
Maps
Climate
Land Use Information System (LUIS)
• Digital Land Use
Map 1:250.000
o ilTTy y
pe
SSoil
pe
ambbiso
isol l
CCam
visool l
LLuuvis
3
RReessuultlt-ID
-ID
11
22
33
44
1
LandUse
Use
Land
Grassland
Grassland
ArableLand
Land
Arable
N2O-emission in
-1 -1
kg N2O-N ha a
1.7
2.1
2.8
3.1
Regional Farm Model
N H 3 - N - V e r lu s t e i n S ta ll
N
BH
e 3t r-iN
e -bVsem
r luos dt ee Lll a g e r
N
InH
t e3n- N
s i- tV鋞e r lu s t e b e i d e r A u s b r in g u n g
SH
t a3llt
N
- Vy ep r lu s t e G e s a m t
KH
鋝3 e- Nr e-Vi e rl u s t e j e h a
N
FH
u t3t e
N
- Nr - V e r lu s t e je G V
N
- Nlain
g e%r d e r N -A u s s c h e id u n g e n
G H黮3le
N
黮3le- N
a -A
u subsr iwnagsucnhgu n g e n je h a
GO
N
D 2eOc -kNu-E
n gms ibs es iio
t rn
a egn je h a
D M
H
z- 躡
G oeftsoarbmilat fnl 鋍
h ee grsec nh u ?t z t
ha
H
Moi lc
ft ohrb
k黨
ila nez- 躡 e rs c h u ? je h a
S tc k .
en
C
F鋜
H 4s- A
us s c he i d ung en K 黨 e N E U S tc k .
NH
a 4c -hVzeurcluhstt e L a g e r N E U
S tc k .
C
i e is
h se iionnheeni tin
e sng/ehsaa m t N E U G V / h a
GH
r o4 遶
C
-E m
Z uHk4a- E
u fmKi sr s
a io
f t fnu et tn
e rjejeh K
d t/S tc k
C
a uh
Z uOk2a- E
um
f Kisrsai o
f tnf ue tnt earuG
m rt E n edrgt i e
C
s feoss as ile
N r p r o d u k tigo nN / h a
Z uOk2a- E
um
f Mis isni o
e nr ea nldK黱r agftefurt-te
C
k gO 2N- E/ m
S is
c hs inoint te n M i n e ra ld 黱 g e r p rkogd u/ hk ati o n
C
NW
- APu s2 s0 cShuemi dmuen g
G
Ne n
u G esa m t kg
NW
- APu s1 s0 c0 hSeui m
d umneg N
en
kg
G
e uje K u h
(Q
uu
e ll
r ,gHe .-U
n j e. 1G9V9 8 k) g
N -A
se
s :c hMe 黮
i d lue n
B e t r i eBbe1tr ie b 1B e t rBi eebt ri 2e b 2 B e Bt r ei et ribe 3
b 3
B Be er irie
e bb 44
B e trt rie
i ebb 55
kg
3 5 0 ha
2 4 0 1, 0712 5 h a1 1 0 9 ,9136 0 5 h a8 0 9 ,4 46 1 6 h a2 3 9 , 9 4 8 7 5 1h8a2 2 ,8 9
kg
A 1 a 6 e i 1 39 4A
, 224 a 6 e i 1 8 2 3 ,1A5 3 a 5 e 1r 13 2 9 ,5 7A 4 a 5 e3r914 , 1 2 A 5 L 6 2e 0i 16 ,9 7
r t, 0 6 In t e n3s9i v7 4 ,6 7
k In
g t e n s i v 1 5 8 4 In
, 1t0e n s i v 6 1 2 9 ,1R9 e d u z i3e9r 8t 9 ,7 9R e d u1z4i e9 1
A n b i n d e22 1 8A
kg
, 3n5b i n d e9 0 6 2 ,2A7 n b i n d6e1 2 8 ,7 9A n b i n2d1 e
2 5 , 1 2 L a u f6s 0t a0ll4 ,5 3
k Jg a/ h a
6 4 J, 6a2
6 1 , 4J7a
5 8 ,2 8J a
6 0 ,0 7 J a
7 1 ,3 9
g e/GuV
3 0H
, 1e0 u
2 6 , 5H9 e u
2 5 ,9 0H e u
3 0 ,2 9 H e u
3 2 ,5 9
kH
%O f f e n
0O
, 3 f5f e n
0 ,3O6 f f e n
0 ,4 0O f f e n
0 , 4 0 G e s c h l. 0 ,3 8
kP
g /rhaallt e lle r 5 7 P
, 2r8a llt e lle r 5 7 , 2P9 r a llt e lle r1 1 ,9 4P r a llt e lle
1 1r , 9 4 P r a llt e lle
5 7r,2 9
kg
,807
1 /6h5a 0 4 3 , 2 6 4 , 8713 2 7 8 4 , 5 14 ,7 44 2 9 8 3 6 , 01 4,9 2 1 2 9 5 2 92, 8, 084 4 0 6 9 9 74, 1
kg N
352
, 0903 5 , 9 3 1 71 1
2 7, 5507 ,5 4
1 6 0 6, 57 0,5 0
36 213, 6, 609
7 48677, 5
,406
k g / h a 5 0 , 4 08 3 , 8 8 2 3 3 , 0668 , 1 6
1 6 9 , 90 6,4 2
5 05, 3, 285
1 2855, 9
,394
4 2 , 5 28 8 , 5 7 1 9 6 , 6942 , 1 8
1 4 3 8, 45 0,8 4
4 820, 5, 211
1 0868, 3
kg
,509
91
, 4358 8 , 6 9
45
3 5, 7405 ,3 9
7 6
6 21
kg
3 31 8, 84 ,4
1 1 896, 4, 659
3 42831, ,7
27
, 1059 0 , 7 5
2 8, 3317 ,5 0
1 92
kg
31
2 0 82 7, 20 5,0 6
6 3 716, 9, 481
1 7 7 325, ,3
8 , 1280 2 , 5 9
1 0 1, 6864 , 5 7
2 49
k g /h a
19 3, 80 8,0 3
1 073, 6, 571
2 082, ,6
5 9 63 9
, 5383 3 , 3 6 3 91 26 7
9 9, 4540 ,8 4
31 05 6
8 42 3, 40 4,4 3
19 85 30711, ,6
7 93
kg
7 6 73 930, 4, 459
1 2 71 7
, 0438 8 , 9 8 11 21 5
2 1, 2887 ,3 3
3 25 3, 02 ,0
9 7
0 63
kg
90
1 8 54 331, 2, 708
4 41 03215, ,2
3 41 1
, 0363 8 , 6 2
34 6
8 8, 3700 ,6 3
1 53 9, 26 ,0
0 1
3 99
kg
9
3 91 783, 5, 976
2 8 63651, ,0
6 2 54 73 3
, 8628 2 , 2 22 5119 9 7
8 8, 7614 ,1 9 1152 29 2
8 08 9, 88 2,8 5
5 36 86974, ,5
1 33
451 30 13 739, 3, 630 110 8
92 53 ,91006 0 , 8 7 1 0 89 4
4 02
8 5, 0816 ,2 9
687
6 12 7, 02 4
,2 2
2 3 0 78 039, 8, 901
5 9 7 39050, ,6
8 4 ,9 1
7 3 ,9 3
6 4 ,6 0
7 6 ,0 3
8 5 ,0 3
Figure 2. SLUSA based integration of process orientated
agro-ecosystem models into GIS
Figure 3. AEIS for the North China Plain
Compared to the availability and the access of topographical
and land use data, soil data are much more difficult to get in
China. Several soil maps in a scale of 1:200,000 were digitized.
The problem here was the lack of information about the soil
properties. Therefore, the landtype map 1:500,000 for the North
China Plain was digitized and soil properties for the soil units
derive from a Chinese soil text book.
The Climate Information System (CIS) consists of daily
weather data for the years 1999-2002. The weather data was
purchased via the German Weather Service (DWD) for 9
weather stations in the North China Plain. Thiessen polygons
were then created to provide spatial weather data on a daily
base. Similar approaches were used by Mathews and Knox
(1999).
Digital spatial land use data are available in China on the scales
1:250,000 and 1:100,000 in ArcInfo vector format. The land
use data derive from LandsatTM land use classifications. The
Institute of Remote Sensing Applications (IRSA) of the China
Academy of Sciences in Beijing is responsible for the land use
classifications. For the North China Plain, the data was
obtained in 1:250,000. Additionally, statistical county data for
the same years were obtained.
Finally, the Agricultural Management Information System
(AMIS) was created on the base of statistical county data,
available survey data, data of the research project and from
literature. For each weather region, a specific management for
the wheat/maize rotation is specified.
All the mentioned data are organized in the frame of the AEIS
which is GIS-based. Consequently, it is possible to identify
unique weather-soil-land-use-management systems (agroecotopes). Approximately 240,000 polygons were created
overlaying the spatial soil, land use and weather data. By using
database functionalities within the GIS 285 unique agroecotopes for the considered wheat/maize rotation are identified.
The parameters of these 285 agro-ecotopes are used as input for
the DNDC model in the site mode. The final result of this
modeling process for the North China Plain is listed in Tab.1.
The spatial distribution of e.g. the emission of N2O is presented
in Fig.4.
Nitrous Oxide Emission for 2000 in kg per ha
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
No data
100
5. DISCUSSION AND CONCLUSIONS
Li et al. (2001) model a national inventory of N2O emissions
from arable lands for the whole of China with the processbased DNDC model. The input parameters for the model are
daily minimum and maximum air temperature, daily
precipitation, nitrogen deposition, soil texture, pH, organic
matter content, individual crop areas, N-fertilizer use, livestock
and human populations, tillage management, irrigation
management, planting and harvest dates, and crop residue
management. Most of the agricultural parameters derive from
official census data. The authors model low N2O emissions for
the North China Plain. This does not correspond to the daily
field measurements of summer maize in 2001 which are
presented in this contribution. One explanation could be that
the measurements of 2001 cannot be compared to the model
results of data of 1990. But from process knowledge, high
emissions from summer maize are expected due to high
precipitation, high temperatures and high N-input during the
growing season. Another explanation for the low regional
model results could be that the used data of the agricultural
census which do not reflect the agricultural practices of distinct
crop management. Only average N-fertilizer data per county
are available. Additionally, the quality of the census data has to
be discussed in general. Tso et al. (1998) mention that the
census data are not reliable and there are wide discrepancies e.g.
of about 25 – 40 % of just how much agricultural land are
actually under cultivation. The question therefore is if regional
modeling of agro-ecosystems should be based on knowledge
based agricultural practices for the different crop managements.
This would make detailed agricultural land use maps an
essential tool for regional agro-ecosystem modeling.
Consequently, RS would become a key tool to provide the
basics to generate the input parameters for regional agroecosystem modeling. Finally, using census data as input
parameters for regional modeling the impact of this data source
should carefully be discussed.
Table 2. Annual N2O-emission, NH3-volatilization and
NO3--leaching for the summer maize/winter wheat area of the
North China Plain (IPCC)
0
Year
N2O
1995
2000
61
62
NO3Gg N yr-1
433
1300
438
1313
NH3
Total
1794
1813
Table 3. Annual N2O-emission, NH3-volatilization and NO3-leaching for the summer maize/winter wheat area of the North
China Plain (SLUSA)
100 Kilometers
Data Source: MoA, 1995; CIESIN, 1997
Map Creation & Layout: Georg Bareth, 2003
Figure 4. N2O emission from summer maize/winter
wheat areas in the North China Plain
Table 1. Annual N2O-emission, NH3-volatilization and
NO3-- leaching for the summer maize/winter wheat area
of the North China Plain (DNDC model)
Year
N2O
1995
2000
52
53
NO3Gg N yr-1
1358
426
1373
429
NH3
Total
1837
1855
Year
N2O
1995
2000
27
27
NO3Gg N yr-1
975
960
985
970
NH3
Total
1962
1982
Additional results of regional modeling of annual N2Oemission, NH3-volatilization und NO3--leaching are presented
in Tab.2 and Tab.3. Different modeling approaches were used.
In Tab.2 the results of the empirical method for national
greenhouse gas inventories of the International Panel on
Climate Change (IPCC) are presented. This empirical method
bases of the application of given factors usually multiplied by
numbers deriving from county statistics. While the total amount
of N losses are very similar to the modeling results of the
DNDC (Tab.1), the numbers for NH3- and NO3--losses are
almost vice versus.
In Tab.3 the results of the GIS- and knowledge-based approach
(SLUSA) are listed. The overall N-losses are again in the same
range of the previous discussed results. But the losses from
N2O-emission, NH3-volatilization and NO3--leaching vary
significantly from the results of the DNDC model and the IPCC
method. Very interesting is the small amount of simulated
N2O-fluxes. Overall, the results of the SLUSA are considered
to be most realistic because regional measurements are
integrated in this approach. Therefore, evaluation and
calibration of the DNDC model and the IPCC method has to be
done on a regional level.
In this contribution, it is clearly shown that the potential of
linking GIS and agro-ecosystem models is huge. Key problems
in this regional modeling approach are still data availability and
evaluation and calibration of the agro-ecosystem models.
Regional AEIS has to be established for general regional
modeling and more regional measurement data has to be
available for these purposes as well as for the assessment of the
simulation results.
ACKNOWLEGEMENT
The project was funded by the German Ministry of Education
and Research (http://www.bmbf.de).
REFERENCE
Bareth, G. (2000): Emissions of greenhouse gases from
agriculture - regional presentation and estimation for a dairy
farm region by using GIS. (In German with English summary).
Hohenheimer Bodenkundliche Studien, H.56.
Bareth, G. and Angenendt, A. (2003): Ökonomischökologische Modellierung von klimarelevanten Emissionen aus
der Landwirtschaft auf regionaler Ebene. Ber. über
Landwirtschaft, Bd.81:29-56.
Bareth, G., Heincke, M. and Glatzel, S. (2001): Soil-land-usesystem approach to estimate nitrous oxide emissions from
agricultural soils. Nutrient Cycling in Agroecosystems,
Vol.60:219-234.
Bareth, G. and Yu, Z. (2004): Availability of digital geodata in
China. (In German with English summary). Peterm. Geogr.
Mitt., Jg.148, H.5:78-85.
Bareth, G. and Yu, Z. (2002): Agri-Enviro-Information-System
for the North China Plain. Proceedings of International Society
for Phozogrammetry and Remote Sensing (ISPRS)
Commission II Symposium of Integrated Systems for Spatial
Data Production, Custodian and Decision Support, 20-23
August 2002, Xian:23-29.
Bareth, G., Yu, Z. and Doluschitz, R. (2002). Set-up of a
Regional Agricultural Environmental Information System
(RAEIS) in China. Proceedings of the EnviroInfo 2002, 25-27
September 2002, Vienna:233-240.
Bill, R. (1999). Grundlagen in GIS (In German), Band 1.
Wichmann: Heidelberg, Germany.
Brown, L., Syed, B., Jarvis, S.C., Sneath, R.W., Phillips, V.R.,
Goulding, K.W.T., and Li, C. (2002): Development and
application of a mechanistic model to estimate emission of
nitrous oxide from UK agriculture. Atmospheric Environment,
Vol.36:917-928.
Hartkamp, A.D., White, J.W. and Hoogenboom, G. (1999):
Interfacing GIS with agronomic modeling: A Review.
Agronomy Journal 91:761-772.
Huber, S., Bareth, G., and Doluschitz, R. (2002): Integrating
the process-based simulation model DNDC into GIS. In W.
Pillmann and K. Tochtermann (editors): Environmental
Communication in the Information Society, Proc. EnviroInfo
Vienna 2002, International Society for Environmental
protection: Vienna:649-654.
Li, C. (2000): Modeling trace gas emissions from agricultural
ecosystems. Nutrient Cycling in Agroecosystems, Vol. 58:259276.
Li, C., Zhuang, Y., Cao, M., Crill, P., Dai, Z., Frolking, S.,
Moore, B., Salas, W., Song, W., and Wang, X. (2001):
Comparing a process-based agro-ecosystem model to the IPCC
methodology for deceloping a national inventory of N2O
emissions from arable lands in China. –Nutrient Cycling in
Agroecosystems, Vol.60:159-175.
Li, C., Frolking, S., Crocker, G.J., Grace, P.R., Klir, J.,
Korchens, M., and Poulton, P.R. (1997): Simulating trends in
soil organic carbon in long-term experiments using the DNDC
model. Geoderma, Vol.81:45-60.
Li, C., Narayanan, V., and Harriss, R., (1996): Model estimates
of nitrous oxide emissions from agricultural lands in the Unites
States. Global Biogeochemical Cycles, Vol.10:297-306.
Liang, Y.L. and Xu, B.C. (2000): Simulated carbon and
nitrogen contents in arid farmland ecosystem in China using
denitrification-decomposition model. Communications in Soil
Science and Plant Analysis, Vol.31:2445-2456.
Matthews, R. and Knox, J. (1999): Up-scaling of methane
emission - experimental results: final report. Dept. Natural
Resources Management, Cranfield University, Silsoe, UK.
Matson, P.A. and Vitousek, P.M. (1990): Ecosystem approach
to a global nitrous oxide budget. Bio Science, Vol.9:667–672.
Plant, R.A.J. (1998): GIS-based extrapolation of land userelated nitrous oxide flux in the atlantic zone of Costa Rica.
Water, Air, and Soil Pollution, Vol.105, pp.131-141.
Seppelt,
K.
(2003):
Computer-based
environmental
management. Wiley-VCH: Weinheim, Germany.
Werner, C. (2003): Erstellung eines N-SpurengasEmissionskatasters für land- und forstwirtschaftlich genutzte
Böden der Bundesrepublik Deutschland. Master Thesis, Inst. f.
Geographie,
Bayerische
Julius-Maximilians-Universität
Würzburg.
Wright, J.R. (1993): GIS and spatial modeling. In J.R. Wright,
L.L.Wiggins, R.K. Jain & T.J. Kim (editors): Expert systems in
environmental planning. Springer: Berlin:83–84.