=Paper=
{{Paper
|id=Vol-1152/paper22
|storemode=property
|title=Information Architecture for Crop Growth Simulation Model Applications
|pdfUrl=https://ceur-ws.org/Vol-1152/paper22.pdf
|volume=Vol-1152
|dblpUrl=https://dblp.org/rec/conf/haicta/FernandesLPT11
}}
==Information Architecture for Crop Growth Simulation Model Applications==
https://ceur-ws.org/Vol-1152/paper22.pdf
Information architecture for crop growth simulation
model applications
J. M. Fernandes1, A. Lazzaretti2, W. Pavan3, and R. Y. Tsukahara4
1
Embrapa Trigo, Passo Fundo, RS, 99001-970, Brazil, e-mail: mauricio@cnpt.embrapa.br
2
IFSUL, Passo Fundo, RS, 99064-440, Brazil e-mail:
alexandre.lazzaretti@passofundo.ifsul.edu.br
3
ICEG, Univ. de Passo Fundo, 99001-970, Passo Fundo, RS, Brazil, e-mail: pavan@upf.br
4
Fundação ABC, Castro, PR, 84165-700,Brazil, e-mail: rodrigo@fundacaoabc.org.br
Abstract. There is a growing concern that climate changes will lead to
significant impact to agriculture, thus having a global effect in food
production. The application of crop models to study the potential impact of
climate change and climate variability provides a direct link between models,
science and the concerns of the society. The objective of this work is to
demonstrate an information architecture that consistently and coherently
achieves the requirements for data availability and use model integration for
crop models researchers. The system includes a database structure for storage
of model input and output data. The information architecture, and its
constituent crop models, is being applied in evaluating farming practices,
technological innovation and climate variability/change impact on agriculture
in Southern part of Brazil. This effort is part of the global Agricultural Model
Intercomparison and Improvement Project (AgMip).
Keywords: climate change, technological innovation, food production,
adaptation.
1 Introduction
Agricultural productivity in both developing and developed countries will have to
improve to achieve substantial increases in food production by 2050 while land and
water resources become less abundant and the effects of climate change introduce
much uncertainty (Antle, 2009). There is a growing concern that the security and
quality of global food production may be affected at large and local spatial scales by
future climate and weather. Adaptation to climate change through changes in farming
practices, cropping patterns, and use of new technologies will help to ease the impact
(Huang et al., 2011).
In the last two decades, substantial efforts have been directed toward
understanding climate change impacts on agricultural systems. The resulting
advances in our understanding of climate impacts have come from the collection of
better data and the observation of actual changes in climate and its impacts. Such
knowledge is critical as we contemplate the use of crop models to design innovative
_________________________________
Copyright ©by the paper’s authors. Copying permitted only for private and academic purposes.
In: M. Salampasis, A. Matopoulos (eds.): Proceedings of the International Conference on Information
and Communication Technologies
for Sustainable Agri-production and Environment (HAICTA 2011), Skiathos, 8-11 September, 2011.
251
technologies and policies to mitigate climate change and facilitate adaptation to the
changes that now appear inevitable in the next several decades and beyond.
The application of crop models to study the potential impact of climate change
and climate variability provides a direct link between models, science and the
concerns of the society. Changing markets, technological innovation and
organizational progress in recent years have increased the intensity and scale of
agricultural land use.
The objective of this work is to demonstrate an information architecture that
consistently and coherently achieves the requirements for data availability and use
and model integration for crop models researchers. Here, we define information
architecture as the structural design of an information space to facilitate task
completion and intuitive access to content. The system includes a database structure
for storage of model input and output data. The constituent crop models within the
information architecture are being applied in evaluating farming practices,
technological innovation and climate variability/change impact on agriculture in
Southern part of Brazil. This data interface enables users to reproduce crop model
simulations, to modify and re-simulate scenarios, and also serves as an archive. The
system also provides the needed tools for databases management. A case study for
the assessment of the impact of climate variability/change on crops such as soybean
and wheat in the state of Paraná will be used to illustrate data flows between weather,
crop models and to effectively perform analyses and present results. Finally, this
effort is part of the global Agricultural Model Intercomparison and Improvement
Project (AgMip) aiming to improve substantially the characterization of risk of
hunger and world food security due to climate change and to enhance adaptation
capacity in both developing and developed countries.
2 Material and Methods
2.1 Crop models overview
The CSM-CROPGRO: Soybean model and the CSM-CERES-CROPSIM: Wheat
model contained within DSSAT (Jones et al., 2003) simulate plant growth and
development from sowing to maturity using a daily time step, and ultimately predict
yield. The physiological processes that are simulated characterize the crop's response
to the major weather factors, including temperature, precipitation and solar radiation,
and to soil characterizations such as the amount of extractable soil water and
nutrients. Daily photosynthesis is a function of light interception and the pool of
carbohydrates available for growth is estimated by daily maintenance and growth
respiration. The remaining carbohydrates are partitioned to vegetative and
reproductive growth as a function of the developmental phase (Boote et al., 1998).
The soil water balance is calculated on a daily basis and is a function of precipitation,
irrigation, transpiration, soil evaporation, and runoff from the soil surface and
drainage from the bottom of the profile. The user distributes soil water among
different horizontal soil layers with depth increments specified. The water content for
252
any soil layer can decrease by soil evaporation, root absorption or flow to an adjacent
layer (Faria and Bowen, 2003). Water stress causes a reduction in photosynthesis and
canopy development, a change in partitioning of biomass and an increase in
senescence or abscission of plant material, depending on the timing and severity of
the stress. Biotic stresses such as those caused by foliar disease causes a reduction in
available photosynthetic tissue and photosynthesis efficiency (Pavan and Fernandes,
2009).
2.2 Climate and soil data
Daily records of weather data are available from Parana Agronomic Institute
(IAPAR) for 28 locations within different agroecological zones (Figure. 1) covering
the state of Paraná, Brazil. For all locations historical records cover 30 or more years
of observations. Changes observed between the period 1980-2009 in temperature and
rainfall were used to develop a future climatic scenario using a weather generator.
The stochastic weather generator can be used for the simulation of weather data at
single sites. Required input data are daily time-series of precipitation, maximum and
minimum temperature, relative humidity, wind speed and solar radiation. The
weather generator calculates a set of statistical properties of historical observations,
creates empirical distributions and generates daily weather datasets. Similarly,
GCM’s (General Circulation Model) outputs from the IPCC Data Distribution
Center, for example, can promptly be inserted in the database. For all locations
which weather was recorded the soil profile data was also available. Soil data
includes soil classification, surface slope, soil color, permeability, and drainage class.
The whole state of Paraná was divided into agroecological zones. Therefore, crop
simulation focused on indicated zones for growing wheat or soybean.
Fig. 1. Map of Paraná state showing the agroecological zones and locations (dots) with soil
and climate data to serve as input to crop models.
253
2.3 Data management
The proposed study considers two approaches for simulating the effect of climate on
wheat and soybean growth and development in the state of Paraná. First, the
estimated potential impacts of past climate on agricultural productivity can be
examined by driving wheat and soybean simulation models with observed weather
data. Secondly the output from a stochastic weather generator can be used in order to
produce climate change scenarios, which are suitable for use in agricultural impact
assessment. In both, crop models can be integrated into geographic information
systems with past and future climate data, to generate map layers representing
interannually variability on phenological dates, yield, disease intensity among others.
The architecture we propose envisions a system designed and implemented on the
top of the PostgreSQL database management system. PostgreSQL extends the
relational data model with support for complex objects, and allows users to add new
programming languages (procedural languages - PL) to be available for writing
special functions and procedures. For the architecture proposed, among other
technologies, we implemented R (pl/r) to sort, analyze and visualize data from crop
model runs. In Figure 2, we present a general a schematic representation of system
architecture that manipulates crop model’s input and output data, making available
through generic web services, implemented to access the database and dynamically
assemble data as requested.
Fig. 2. Database management and exchange system schematic (Adapted from AgMip, 2011).
2.4 Data analysis
The information architecture evaluates performance of CSM-CERES-CROPSIM:
Wheat and CSM-CROPGRO: Soybean model by calculating the Pearson correlation
254
coefficient (r) and root mean squared error (RMSE) between the modeled and the
corresponding observed yield series at both the crop model local scale and state
scale. Furthermore, temporal and spatial changes of wheat and soybean productivity
can be presented in a probabilistic framework, based on simulation outputs. For
example, probability density functions (PDFs) of wheat and soybean yield changes
during 2030s, 2050s, and 2080s, respectively, relative to 1980–2009, at the
representative locality. Across all the wheat and soybean cultivation regions in the
state of Paraná, the system is also programmed to derive histogram and cumulative
distribution function (CDF) of wheat and soybean yields during 1980–2009, and
wheat and soybean yield changes during 2030s, 2050s, and 2080s, respectively.
Finally, the system can plot the spatial changes of mean and standard deviations of
wheat and soybean productivity during 1980–2009, and their changes during 2030s,
2050s, and 2080s, across the study region.
3. Results
A dynamically build webpage provides access to figures and code from the
simulation experiment. It can be viewed with any standards compliant browser with
Javascript and CSS support enabled. User can click on the treatments links for
treatment details and a listing of links for corresponding plots (Figure 3).
Lattice (Sarkar, 2008), which is an R data visualization package, proved to be very
useful for handling multivariate simulation output data. The attraction to lattice is
derived from wide variety of displayable graphics, its portability as it is written
entirely in R. Customizability is owing to a variety of graphical parameters, flexible
use of panel functions, extensibility through the object model and its leverage of the
underlying R grid graphics system (R Development Core Team 2011) upon which it
is built. The nature of lattice is shown in the multipanel graphic (Figure 4.). Here,
yield is numeric while Agroecological Zones, Variety and Sowing Dates are factors.
These illustrated the simulation of soybean yield in two contrasting agroecological
zones. Soybean crop was simulated using three sowing dates with observed weather
data during the period of 1980 through 2009.
Hopefully, this information system will prove to be worthwhile in the assessment
of impact of climate variability/change in agriculture across a wide region. Thus,
enhancing capabilities in the developed and developing world for responding to
climate change by building scientific and technical capacity, advancing scientific
knowledge, and linking scientific and policy communities.
255
Fig. 3. Webpage screen capture showing experiment list, corresponding plots and R code.
256
Fig. 4. Simulated soybean yield (Kg/ha) of three soybean varieties, three sowing dates
(October 1st, November 1st and December 1st ), using observed weather data during 1980-2009
in the agroecological zones of Southern and Northern of Paraná state, Brazil.
Acknowledgments. This work is underway and is being partially supported by
FINEP (Project Number 01.09.0324-00)
References
1. AgMIP (2011). AgMIP - Agricultural Model Intercomparison and Improvement
Project. Information Technologies Team Protocols.
http://www.agmip.org/?page_id=171
2. Antle, J. (2009). Climate change and agriculture: economic impacts. Choices
23:9-11.
3. Boote, K. J., Jones, J. W., Hoogenboom, G. and Pickering, N. E. (1998). The
CROPGRO model for grain legumes. p. 99–128. In G.Y. Tsuji et al. (ed.)
257
Understanding options for agricultural production. Kluwer Academic Publ.,
Dordrecht, the Netherlands.
4. Faria, R. T. and Bowen, W. T. (2003). Evaluation of DSSAT soil-water balance
module under cropped and bare soil conditions. Braz. arch. biol. Technol, 46(4):
489-498 .
5. Huang, H., Lampe, M. von, Tongeren, F. van. (2011). Climate change and trade
in agriculture. Food Policy, 36(1), The challenge of global food sustainability. p.
S9-S13.
6. Jones, J. W., Hogemboom, G., Porter, C. H., Boote, K. J., Batchelor, W. D.;
Hunt, L. A., Wilkens, P. W., Singh, U., Gijsman, A. J. and Ritchie, J. T. (2003)
The DSSAT cropping system model. European Journal of Agronomy, 18: 235-
265.
7. Pavan, W. and Fernandes, J. M. C. (2009). Uso de orientação a objetos no
desenvolvimento de modelos de simulação de doenças de plantas genéricos.
Revista Brasileira de Agroinformática, 9: 12-27.
8. R Development Core Team (2011). R: A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-
900051-07-0, URL http://www.R-project.org/.
9. Sarkar, D. (2008). Lattice: Multivariate Data Visualization with R. Springer, New
York. ISBN 978-0-387-75968-5.
258