Characterizing Temporal Dynamic of Weather Variability to Support Decision Making on Weed Control
CHARACTERIZING TEMPORAL DYNAMIC OF WEATHER
VARIABILITY TO SUPPORT DECISION MAKING
ON WEED CONTROL
RIZKY MULYA SAMPURNO
GRADUATE SCHOOL
BOGOR AGRICULTURAL UNIVERSITY
BOGOR
2014
STATEMENT
I, Rizky Mulya Sampurno, hereby declare that this thesis entitled
Characterizing Temporal Dynamic of Weather Variability to Support Decision
Making on Weed Control
Is a result of my work under the supervision advisory board and that it has not been
published before. The content of the thesis has been examined by the advisory board
and external examiner.
Bogor, February 2014
Rizky Mulya Sampurno
G051110091
RINGKASAN
RIZKY MULYA SAMPURNO. Karakterisasi Dinamika Temporal Variabel
Cuaca untuk Mendukung Pengambilan Keputusan pada Pengendalian Gulma.
Dibimbing oleh KUDANG BORO SEMINAR dan YULI SUHARNOTO.
Pada pertanian modern, herbisida dominan digunakan untuk pengendalian
gulma. Meskipun herbisida memiliki manfaat positif dalam memberantas gulma,
herbisida juga berpotensi dapat mencemari lingkungan apabila terjadi spray drift
pada saat penyemprotan. Spray drift dapat terjadi karena kondisi cuaca yang tidak
sesuai. Pengetahuan tentang kondisi cuaca akan membantu petani dan pengambil
keputusan untuk menentukan teknologi dan metode tepat guna dalam
memberantas gulma serta meminimalkan spray drift dan pemborosan herbisida
lainnya.
Kemajuan teknologi informasi telah diterapkan secara luas di bidang
pertanian seperti halnya pertanian presisi. Solahudin (2013) mengembangkan
metode pengendalian gulma berbasis multi-agen. Sistem tersebut memiliki dua
fungsi yaitu fungsi konsultasi dengan memberikan rekomendasi sebelum
melakukan penyemprotan (off-farm) dan pengendalian selama aplikasi
penyemprotan (on-farm). Sistem ini telah diterapkan dalam pengendalian gulma
tanaman kacang tanah.
Hubungan dengan penelitian ini adalah sistem yang dibangun oleh
Solahudin (2013) perlu ditingkatkan kemampuannya dalam hal yang berhubungan
dengan data cuaca. Data cuaca dapat digunakan untuk mengoptimalkan jadwal
penyemprotan pada kegiatan pengendalian gulma yang ramah lingkungan serta
untuk menyiapkan peralatan dan mesin yang akan digunakan sebelum hari
penyemprotan. Kondisi cuaca baik spasial maupun temporal merupakan informasi
yang penting bagi kegiatan pertanian. Integrasi satelit meteorologi dengan produk
prediksi cuaca numeric cukup menjanjikan untuk digunakan sebagai masukan
pengambilan keputusan dalam aplikasi penyemprotan terutama untuk daerah yang
kurang terjangkau oleh stasiun cuaca.
Tujuan dari penelitian ini adalah untuk mengembangkan sistem pendukung
keputusan (DSS) untuk penjadwalan pengendalian gulma dan untuk memilih
ukuran nozzle semprot yang dapat meminimalkan dampak negatif terhadap
lingkungan. Data utama yang diperlukan untuk sistem yang diusulkan ini meliputi
data cuaca 10 tahun yang diperoleh dari satelit penginderaan jauh seperti Oceanic
& Atmospheric Administration Nasional (NOAA), Tropical Rainfall Measuring
Mission (TRMM) dan data indeks vegetasi dari MODerate Resolution Imaging
Spectroradiometer (MODIS). Data indeks vegetasi digunakan untuk menentukan
periode tanam padi. Sedangkan data cuaca digunakan untuk menentukan jadwal
penyemprotan dan untuk menentukan ukuran nozzle semprot yang tepat
berdasarkan kondisi cuaca.
Dalam pertanian presisi, pengendalian gulma dilakukan dua kali yaitu preplanting dan post-emergence. Untuk mengetahui periode ini, kami menggunakan
waktu tanam sebagai referensi. Kami menggunakan tanaman padi untuk penelitian
ini. Waktu tanam tanaman padi mudah diidentifikasi melalui analisis multi
temporal indeks vegetasi yang diperoleh dari penginderaan jauh. Indeks vegetasi
EVI padi pada areal sawah di daerah penelitian dapat menunjukkan siklus
pertumbuhan padi tahunan. Berdasarkan analisis dari EVI, diperkirakan bahwa
waktu penyemprotan padi adalah pada bulan April-Mei untuk siklus tanam 1dan
Oktober-November untuk siklus tanam 2.
Karakteristik pola cuaca di Jonggol diamati selama sepuluh tahun. Setiap
parameter cuaca memiliki karakteristik tersendiri dan umumnya memiliki pola
yang sama dar tahun ke tahun. Pada umumnya, selama sepuluh tahun curah hujan
yang tinggi terjadi pada akhir tahun sampai dengan awal tahun. Sedangkan curah
hujan rendah pada pertengahan tahun. Kecepatan angin berfluktuasi. Angin yang
tinggi terjadi pada akhir tahun sampai Januari dengan kecepatan lebih dari 10
km/jam. Selama sepuluh tahun, suhu minimum dan suhu maksimum adalah
23.57°C dan 30.5°C. Kelembaban relatif menurun bersamaan dengan kenaikan
suhu udara. Kelembaban berkisar antara 67.5 – 95.5%. Kelembaban tinggi terjadi
pada akhir tahun sampai awal tahun, dan kelembaban relatif menurun pada
pertengahan tahun. Petani dan pengambil keputusan dapat menggunakan
informasi cuaca tahun sebelumnya untuk mengetahui waktu yang optimal untuk
penjadwalan aplikasi penyemprotan, untuk mempersiapkan alat dan mesin
termasuk sprayer. Waktu yang optimal untuk pengendalian gulma ditentukan dari
interval waktu penyemprotan yang diperoleh dari analisis indeks vegetasi dengan
mempertimbangkan kondisi cuaca.
Aplikasi untuk menentukan ukuran nozzle semprot telah dikembangkan
berdasarkan kriteria-kriteria kondisi cuaca. Pengetahuan untuk menentukan
ukuran nozzle diperoleh dari penelitian sebelumnya. Curah hujan dijadikan
sebagai parameter utama dalam menentukan keputusan untuk melakukan
penyemprotan. Parameter selanjutnya adalah angin, suhu, dan kelembaban.
Parameter cuaca dapat diinput secara manual atau dapat menggunakan data cuaca
yang tersimpan database.
Sistem penunjang keputusan ini telah dikembangkan dan diuji dengan data
riil di lapangan yang diperoleh dari satelit penginderaan jauh. Sistem yang
dikembangkan dapat memberikan
informasi waktu yang optimal untuk
penyemprotan pada tanaman padi serta dapat menentukan ukuran nozzle semprot
berdasarkan kondisi cuaca. Tidak hanya untuk tanaman padi, sistem ini dapat
diterapkan baik pada tanaman di dataran rendah maupun di dataran tinggi.
Kata kunci: DSS, herbisida, spray drift, pola cuaca, pengendalian gulma
SUMMARY
RIZKY MULYA SAMPURNO. Characterizing Temporal Dynamic of Weather
Variability to Support Decision Making on Weed Control. Supervised by
KUDANG BORO SEMINAR and YULI SUHARNOTO.
Herbicide is the dominant tool used for weed control in modern agriculture.
Although herbicide has positive benefit in killing the target weeds, it potential
becomes negative impact to the environment if some remains in the air and drift.
Spray drift can happen due to unsuitable weather. The knowledge of weather
condition will help farmer and decision maker to decide the appropriate
technology and method for eradicating weed which minimize drift and other
potential waste.
The progress of information technology has been applied widely in
agriculture such as precision agriculture. Solahudin (2013) developed a weed
control method using multi-agents based. That system has two functions i.e.
consultation function by giving recomendation before spraying (off-farm) and
controlling during spray application (on-farm) using multi-intelligent agents
which applied for groundnut farming.
Relationship with this research is the system built by Solahudin (2013)
needs to be improved in knowledge that related with weather. Weather data can be
used to optimize spray scheduling on weed control activities that safe to the
environment and for setup the equipment and machinery which prepared earlier
before application day. The spatial and temporal variability weather conditions are
important sources for agricultural activities such as spray application. Integration
meteorological satellite with numerical weather prediction (NWP) product is
promising in find timely weather variables as input for decision making to resolve
problems in spray application especially for area which sparse coverage of
weather stations.
The objective of this research is to develop a decision support system (DSS)
for schedulling of weed control and for selecting the proper nozzle size of the
sprayers that introduce minimum negative impact to the environment. The main
set of data required for our proposed system includes the set of 10 years weather
data series acquired from remote sensing such as the National Oceanic &
Atmospheric Administration (NOAA) and the Tropical Rainfall Measuring
Mission (TRMM) and a set of vegetation index from the MODerate Resolution
Imaging Spectroradiometer (MODIS). The vegetation index data utilized to
determine the planting period of paddy and weather data set utilized to determine
spray schedule and to determine the proper size of the sprayers for weed control.
In precision farming, weed control is done two times i.e. pre-planting and
post-emergence. To know these times, we used planting time as reference. We
used paddy for this study. Paddy planting time is easy to identify through multitemporal analysis of vegetation index. Enhanced vegetation index (EVI) of paddy
field in study area shows the annual paddy growth cycle, it representing intensive
cropping with multiple harvests. Based on analysis of EVI, we estimated the
spraying times for paddy are April to May and October to November for cycle 1
and cycle 2, respectively.
We study weather pattern in Jonggol during ten years. Every parameter have
own characteristic and generally in same fluctuated pattern form. Generally,
during ten years rainfall is high in year end to early year while low in mid-year.
Wind speed is fluctuates. Wind is high in year end to January every year about
more than 10 km/s. For ten years, minimum and maximum temperatures are
23.57°C and 30.5°C. Relative humidity decreased when air temperature increase.
It is about 67.5 – 95.5%, high in year end to early year, and lowers in middle year.
Farmer or decision maker can use information from past weather data to find out
optimal time for scheduling, preparing machinery and sprayer. Optimal week for
weed control determined from interval time for spraying both in crop cycle 1 and
crop cycle 2.
We developed application to determine the proper nozzle size for sprayer
based on weather condition. Knowledge to determine nozzle size is acquired from
previous research. Rainfall is the first parameter which decides do spray or do not
spray, because spray application will not conducted in rainy day and herbicide
particles will run off along with rain water. Wind becomes second parameter,
following by temperature and humidity. Weather parameters can be inputted
manually or can use weather data taken from the past datasets which stored in
database.
The DSS for weed applications has been developed and tested with a real
data set acquired from remote sensing devices. The developed system can
generate optimal spray scheduling and recommend the proper size of nozzles used
for spray application on paddy crops based on the weather condition, and thus
minimizing spray drift and bad environmental impact. This method could be
implemented both for low-land crop and high-land crop.
Key words: DSS, herbicides, spray drift, weather pattern, weed control
Copyright © 2014, Bogor Agricultural University
Copyright are protected by law
It is prohibited to cite all of part of this thesis without referring to and mentioning the
source. Citation only permitted for the sake of education, research, scientific writing,
report writing, critical writing or reviewing scientific problem. Citation does not
inflict the name and honor of Bogor Agricultural University
It is prohibited to republish and reproduce all part of this thesis without written
permission from Bogor Agricultural University
CHARACTERIZING TEMPORAL DYNAMIC OF WEATHER
VARIABILITY TO SUPPORT DECISION MAKING
ON WEED CONTROL
RIZKY MULYA SAMPURNO
A thesis submitted for the degree Master of Science in Information
Technology for Natural Resources Management Study Program
GRADUATE SCHOOL
BOGOR AGRICULTURAL UNIVERSITY
BOGOR
2014
The External Examiner:
Dr Liyantono, STP, MAgr
Thesis Title : Characterizing Temporal Dynamic of Weather Variability to
Support Decision Making on Weed Control
Name
: Rizky Mulya Sampurno
Student ID : G051110091
Approved by
Advisory Board
Prof Dr Ir Kudang Boro Seminar, MSc
Supervisor
Dr Ir Yuli Suharnoto, MEng
Co-Supervisor
Endorsed by
Program Coordinator of MSc in IT
for Natural Resources Management
Dr Ir Hartrisari Hardjomidjojo, DEA
Date of Examination:
January, 29th 2014
Dean of Graduate School
Dr Ir Dahrul Syah, MScAgr
Date of Graduation:
Thesis Title : Characterizing Temporal Dynamic of Weather Variability to
Support Decision Making on Weed Control
:
Rizky Mulya Sampumo
Name
Student ID : GOS111 0091
Approved by
Advisory Board
Prof Dr Ir Kudang Boro Seminar, MSc
Supervisor
Dr Ir Yuli Suhamoto, MEng
Co-Supervisor
Endorsed by
Program Coordinator of MSc in IT
for Natural Resources Management
Dr lr Hartrisari Hardjomidjojo, DEA
Date of Examination:
1h
January, 29 2014
Date of Graduation :
28
MAR 2014
ACKNOWLEDGMENTS
Alhamdulillah, first and most, I thank to The Almighty Allah subhanahu wa
ta’ala for allowing me alive and get many experiences in this world. Plenty rahmats
have been given makes me easier to complete this thesis. The topic selected for this
research is utilization of remote sensing technology for agricultural application that
conducted from July – November 2013 with the title Characterizing temporal
dynamic of weather variability to support decision making on weed control.
I owe a great deal to Prof Dr Kudang Boro Seminar who served as chair of the
board of supervisor and teach anything. Dr Yuli Suharnoto as co-supervisor made
significant contribution to many aspects in the thesis, particularly in remote sensing.
Thank you very much Pak Kudang and Pak Yuli for guidance, patience, comments,
correction and constructive inputs through all months of my research.
I would like to thank Dr Mohamad Solahudin for inputs and idea for this
research. My appreciation also goes to Dr Liyantono and Dr Bib Paruhum Silalahi as
the external examiners for their valuable critics, inputs, and corrections. I would like
to thank Dr Ibnu Sofyan, Mr Harry Imantho, Dr Tofael Ahamed, Dr Ryozo Noguchi,
Dr Yudi Setiawan for the valuable inputs and comments. I would like to thank to all
MIT secretariat that support our administration, technical and facility. I would like to
thank Agus Aris and all my colleagues in MIT for helping, supporting, and
togetherness. Special thanks go to Nadia Tannia Hendartina for any help and support
during my study.
This work is dedicated to my parent, Ibu and Bapak who cheer my life at the
beginning and afterward, prayer, financial, encouragement and support. Hopefully,
the results of this research would provide a positive and valuable contribution for
anyone who reads it.
Bogor, February 2014
Rizky Mulya Sampurno
LIST OF CONTENTS
LIST OF TABLES
vii
LIST OF FIGURES
vii
LIST OF APENDICES
vii
LIST OF ABREVIATON
viii
1 INTRODUCTION
Background
Problem Formulation
Objective
Research Benefits
Scope of Research
1
1
2
2
3
3
2 LITERATURE REVIEW
Weed
Weed Control on Paddy Field using Herbicide
Spray Drift
Paddy Phenological Stages
Weed Control in Precision Farming
3
3
4
5
5
7
3 METHOD
Study Location
Time of Research
Data
Tools Requirement
Data Processing
Data Analysis
Identification of Paddy Phenology
Estimation of Weed Control Time
Development Application to Determine Nozzle Sprayer
8
8
10
11
12
12
13
13
14
14
4 RESULT AND DISCUSSION
Paddy Planting Time
Weed Control and Spray Application Time
Weather Pattern
Application for Minimizing Spray Drift
Herbicide Application in Agricultural crop
15
16
17
18
21
22
5 CONCLUSION AND RECOMMENDATION
Conclusion
Recommendation
23
23
23
REFERENCES
23
APPENDICES
27
CURRICULLUM VITAE
34
LIST OF TABLES
1
2
3
EVI distribution of paddy
List of datasets of research
Decision making to determine nozzle size based on weather parameter
6
12
15
LIST OF FIGURES
1
2
3
4
5
6
7
8
9
10
11
12
Weed on some main crops; (a) paddy, (b) groundnut, and (c) maize
Architecture of supervisory control system based on precision farming
(Solahudin 2013)
Location of study, Jonggol, Bogor district, West Java, Indonesia
Representation of paddy field area in MODIS Image (One pixel = 500
x 500 m)
Examples of paddy field area in one pixel 500 x 500 m
(www.bing.com/maps)
General framework of study
Utilization of remote sensing data for weed control
Weed control time which estimated from planting time through
enhanced vegetation index (EVI)
Weather patterns of study area and it utilization to characterize
weather condition during spray duration
Weather condition during spray application interval in both paddy
planting cycle 1 period March – May (a) and paddy planting cycle 2
period September – November (b)
Knowledge representation of nozzle selection based on weather
condition; P: rainfall, W: wind speed, T: temperature, and RH: relative
humidity
Sprayer nozzle selection based on normal weather condition, (a)
application in normal condition and (b) application in extreme
condition
3
8
9
9
10
11
16
17
18
20
21
22
LIST OF APPENDICES
1
2
3
4
Coordinate of several paddy fields in Jonggol
27
Enhanced vegetation index (EVI) value of several paddy fields in Jonggol 28
Examples of processing netCDF data
29
Sourcecode of Application for Minimizing Spray Drift
31
LIST OF ABBREVIATION
BPS
cdo
DAP
DSS
EVI
ESRL
GeoTIFF
GIS
GPI
I/O
LAPAN
LP DAAC
LSWI
MODIS
NCEP
NDVI
NDWI
netCDF
NOAA
NWP
PR
PSD
RH
SDLC
TMI
TRMM
USGS
VIRS
WV
Badan Pusat Statistik
Climate Data Operator
Days After Planting
Decision Support System
Enhanced Vegetation Index
Earth System Research Laboratory
Geostationary Earth Orbit Tagged Image File Format
Geographic Information System
Global precipitation index
Input/Output
Lembaga Penerbangan dan Antariksa Nasional
Land Processes Distributed Active Archive Center
Land Surface Water Index
MODerate Resolution Imaging Spectroradiometer
National Centers for Environmental Prediction
Normalized Difference Vegetation Index
Normalized Difference Water Index
Network Common Data Form
National Oceanic & Atmospheric Administration
Numerical Weather Prediction
Precipitation Radar
Physical Sciences Division
Relative Humidity
System Development Life Cycle
TRMM Microwave Image
Tropical Rainfall Measuring Mission
United State Geological Survey
Visible and Infrared Scanner
Wind Velocity
1
1 INTRODUCTION
Background
Weeds are a serious problem for agricultural crop. They rob main crops of
sunlight, water and nutrient causing production losses both in quantity and quality
(Solahudin 2013). Losses due to weed are: wheat (9.8%), rice (10.8%), maize
(13%), sorghum (17.8%), potatoes (4%) and groundnut (11.8%) (Evans 1996).
Even, an uncontrolled weed can decrease yield until 20-80% (Utami 2004). In
modern way, agriculture activity is depends on utilization of chemical substance
to increase production. Herbicides are the dominant tool used for weed control in
modern agriculture (Harker et al. 2013).
Although herbicide has positive benefit in killing the target weeds, it
potentially becomes negative impact if some remains in the air and drift. Spray
drift from herbicide can cause crop protection chemicals to be deposited in
undesirable areas (Nuyttens et al. 2011). It has serious consequences such as
damage to sensitive adjoining crops (Reddy et al. 2010), damage to susceptible
off-target areas (Marrs and Frost 1997), environmental contamination, illegal
herbicide residues, lowers yield results (Kjaer et al. 2006) and health risks to
animals (Otto et al. 2009) and people (Butler et al. 2010).
Spray drift continues to be a major problem in applying herbicides. Factors
that cause drift are weather conditions and sprayer setup (Wolf 2006). Drift can
happen due to unsuitable weather. It potentially occurred every time when sprayer
turned on. The knowledge of weather condition will help farmer and decision
maker to decide the appropriate technology and method for eradicating weed, plan,
and effectively execute spray applications to avoid spray drift and other potential
waste.
The progress of information technology has been applied widely in
agriculture such as precision agriculture (Auernhammer 2001). Solahudin (2013)
has developed a weed control method in precision agriculture using multi-agents
based. The method is a supervisory system to determine technology and liquid
applicator capacity and controlling agents. That system has two functions i.e. as
consultation function by giving recomdation before spraying (off-farm) and
controlling during spraying using multi-intelligent agents (on-farm) which applied
for groundnut farming. Decision making method influence by factors on weeding
activity such as time schedule, total area, working time, type of crop, type of weed,
weed density, herbicide, weather, and sprayer technology. While agents are for:
image acquisition, filtering, crop detection, determination weed density, and
determination herbicide dosage.
The weather conditions both spatially and temporally have not studied more
in previous research. Relationship with this research is the system built by
Solahudin (2013) needs to be improved in knowledge that related with weather.
Weather data can be used to optimize spray scheduling in weed control activities
to avoid drift and environmental friendly. Also it can be used in preparing the
equipment and machinery that determine earlier before application day.
The spatial and temporal variability weather conditions are important
sources for agricultural activities such as spray application. Integration
2
meteorological satellite with the Numerical Weather Prediction (NWP) product is
promising in find timely weather variables as input for decision making to resolve
problems in spray application especially for area which sparse coverage of
weather stations (Bahar 2012; Roebeling et al. 2004; De Wit et al. 2010).
However, the availability of data in real-time is still difficult to achieve. The
Tropical Rainfall Measuring Mission (TRMM) data is capable of providing daily
rainfall. NWP products from NCEP/NOAA such as 2 m temperature, wind, and
RH (Relative Humidity) are used as other input. Moreover data from experience
could be used for scheduling and as decision support for preparing tools and
machinery before spray application conducted.
Problem Formulation
In addition to the effectiveness and efficiency issues, drift of chemical
substance on herbicide application can cause damage to other agricultural product
and pollute the environment. Utilization of weather data to resolve these problems
is needed. However, the availability of data in real-time is still difficult to achieve.
Therefore, utilization of remote sensing data is expected can help to resolve those
problems.
Objective
The objective of this research is to develop a decision support system (DSS)
for schedulling of weed spraying and for selecting the proper nozzle size of the
sprayers that introduce minimum negative impact to the environment.
Research Benefits
The benefits of this research are:
- Introduce a method that utilizes weather data for agriculture application.
- Provide information about characteristics of weather conditions on study area.
- Provide recommendation the proper nozzle size of sprayer based on weather
condition.
Scope of Research
This research is limited to the development of weed control methods based
on precision agriculture. The developed method is functioned as decision support
systems for spray scheduling activities and applications to determine the proper
nozzle size of sprayer based on weather conditions. We employ weather and
vegetation data remotely sensed from satellite observation result. It is also use to
overcome the limitation of meteorology station in providing weather data. This
weed control method suitable to be applied on big scale farming area. The main
crop that observed in this study is paddy. Paddy fields easily identified using
3
remote sensing especially for big scale farming area. Paddy becomes main
agricultural product that planted in many region of Indonesia. This developed
DSS can be applied not only for paddy but also for other crop which used spray
application.
2 LITERATURE REVIEW
Weed
Weeds are unwanted plants growing along with agricultural crops (Dangwal
et al. 2011). Weeds compete with crops mainly for space, sunlight, moisture,
nutrient, and reduce the quantity as well as quality of production (Solahudin
2013). Holm et al. (1979) estimated 250 weed species which are harmful for
agricultural crops throughout the world. The competition of weeds for nutrients
may results in such obvious responses as dwarfing in plant size, nutrient starved
conditions, wilting and actual dying out of plants (Anderson et al. 1996). Weed
seeds germinate earlier to agricultural crops, their seedlings grows faster and
aggressive so that they crowd out all other plants which possesses more valuable
properties and establish a kingdom of their own within a short period of time.
Weed species mature ahead of crops so that their seeds are collected with the crop
harvest and get distributed to other places. Some weed species caused damage to
crops by harboring pests and disease agents.
(a)
(b)
(c)
Figure 1 Weed on some main crops; (a) paddy, (b) groundnut, and (c) maize.
Weeds act as host for bacteria, viruses and nematodes that causes diseases in
crop plants (Peters et al. 1955). Weeds show allelopathic effects on agricultural
crops by secreting allelochemicals that inhibit their growth and germination
(Oudhia P and Tripathi RS 1998). The weedy crop may sometime leads to
complete failure. Weeds cause yield losses as reported by Evan (1996) were wheat
(9.8%), rice (10.8%), maize (13%), sorghum (17.8%), potatoes (4%), sugar
(15.7%), soybean (13.5%), cacao (11.9%) and groundnut (11.8%). An
uncontrolled weed can decrease yield until 20-80% (Utami 2004). The cost of
removing weeds adds to the cost of production of crops, thus producers losses part
of their investment and the country suffers a reduction in agricultural products.
4
In irrigated area, competition between weed and paddy can decrease rice
yield 10-40%, depend on species and density of weed, soil type, water and
weather (Nantasomsaran and Moody 1993). Nyarko and De Datta (1991) reported
that yield loss of paddy by weed can be 10-15%, while Pane et al. (2002) in
Karawang, yield loss about 8-12%. It shows that weed control which perform by
farmer is not effective especially for manual weeding. World Bank (1996)
reported in 1995, weed decrease yield in Asia about 50 million tones. Competition
weed and paddy crop decreased yield 10 million tones every year in China
(Labrada 2003).
Weed Control on Paddy Field using Herbicide
There are many methods of weed control. It can be categorized into two
groups, i.e. indirect method and direct method (Pane and Jatmiko 2009). Indirect
methods are: (1) using good seed and variety, (2) soil tillage, (3) irrigation and
fertilizer control, (4) controlling planting space, (5) crop rotation, and (6)
biological method. Direct methods are: (1) hand pulling, (2) mechanic method,
and (3) herbicide.
In high density population country, the salary of worker is relatively
inexpensive. Weed control perform manually by hand pulling and flooding (Pane
and Jatmiko 2009). Recently, hand pulling is not effective but the worker is
limited. Then, weed control is performed using mechanical method. The next
generation is development of chemical method known as herbicide. It is claimed
more efficient, cheap, quick, and minimal worker.
Nowadays, there are many farmers that use herbicide to control weed (Pane
and Jatmiko 2009). In irrigation paddy field such as in outer Java Island, weeding
worker is rare and expensive. In West Java especially irrigated area of Jatiluhur,
because the planting time is same, all farmers need worker while the worker is
limited. Also in paddy field area near the city, the worker is also limited because
younger worker tend to work in office, factory, property etc.
De Datta and Herdt (1983) reported that in agricultural area where the salary
of worker is expensive and well irrigation management area such as Japan, Korea,
and Taiwan, about 75-100% of farmer have used herbicide. According to (Chisaka
1995), combination of herbicide with indirect methods like soil tillage and
fertilization prove more effective than hand pulling. Direktorat Jendral Pertanian
Tanaman Pangan (1982) reported that farmers in Deli Serdang (North Sumatera),
Musi Banyuasin (South Sumatera), Sidrap (South Celebes), and Karawang and
Indramayu (West Java) used herbicide 21%, 37.5%, 100%, and 17.5%,
respectively.
Each plant species has a different critical period in its competition with
weeds. Generally critical period of weed competition starts since the plant grows
to 1/4-1/3 the first of the plant life cycle (Pane and Jatmiko 2009). For example in
paddy, critical period of weed occurs by age the first 40 days of its life cycle. In
this stage, canopy of paddy is not closed and the sunlight intensity is high in the
ground. Weeds seed germinate and grow faster than paddy.
5
Spray Drift
Spray drift is becoming an increasingly important part of every spraying
operation. Spray drift is the movement of a pesticide/herbicide through the air,
during or after application, to a site other than the intended target. Drift is
considered to be the most challenging problem facing applicators and
pesticide/herbicide manufacturers. Although drift may occur as vaporized active
chemical substances from the application site, it is usually the physical movement
of very small drops from the target area at the time of application (Hofman and
Elton 2001).
Spray drift and from agricultural pesticides/herbicides can cause crop
protection chemicals to be deposited in undesirable areas (Nuyttens et al. 2011).
This can have serious consequences such as damage to sensitive adjoining crops
(Reddy et al. 2010), damage to susceptible off-target areas (Marrs and Frost
1997), environmental contamination, illegal pesticide/herbicide residues, lower
yield results (Kjaer et al. 2006) and health risks to animals (Otto et al. 2009) and
people (Butler et al. 2010).
There are several factors that play a significant part in the creation and
reduction of drift (Anonymous 2010; Hofman and Elton 2001). They can be
grouped into one of the following categories: (a) spray characteristics, such as
volatility and viscosity of the pesticide formulation, (b) equipment and application
techniques such as nozzle sprayer and distance between target weed with nozzle
sprayer, (c) weather conditions at the time of application (wind speed and
direction, temperature, relative humidity and stability of air at the application
site), (d) operator care, attitude and skill.
Paddy Phenological Stages
The phenology of paddy is divided into several stages (Semedi, 2012). The
first stage is flooded. It performs before the paddy seed being planted. Once the
seed planted or at germination stage, it will take 25-30 days to reach tillering and
the paddy leaf starting to grow. The leaf color then turning from yellow into green
in tillering stages. The growth will reach maximum at the panicle where for early
variety of paddy it will take 55-60 days from germination stage and for late
variety it will take 65-75 days. It is called heading time. The next stage is the
flowering, paddy started to grow flowers and the leaf color started to turn into
yellow. For the early variety of paddy, this stage reaches in 85-100 days after
germination and for the late variety of paddy it needs 100-115 days. The last stage
is harvesting, where paddy is already developed and ready to be harvested. The
early variety of paddy needs 130-145 days after germination to be harvested,
while for the late variety needs 140-165 days. The paddy phenological stages can
be studied using vegetation index.
A number of researches has been attempted to detect paddy phenological
stages in wide extent of Asian region using MODIS data (Xiao et al. 2005; 2006,
Sakamoto et al. 2005, Uchida 2007, Sari et al. 2010 and Setiawan et al. 2011.
They employed a kind of indices derived from MODIS data such as Enhanced
Vegetation Index (EVI), Normalized Difference Vegetation Index (NDVI), LSWI,
6
and NDWI. Sakamoto et al. (2005) and Uchida (2007) used time-series EVI data
to determine the planting date, heading date, harvesting date, and growing period
of paddy. Sari et al. (2010) study paddy phenology in paddy fields with complex
cropping pattern.
Sari et al. (2010) utilized MODIS EVI to detect growing stages of paddy
crop cycle. Sari et al. (2010) characterized the phenological stages of paddy into
three main periods: (1) the flooding and transplanting period; (2) the growing
period (vegetative growth, reproductive and ripening stages); and (3) the fallow
period after harvest. Acording to Domiri et al. (2005), during flooding and paddy
transplanting period, about 0-20 days after transplanting (DAT), EVI values < 0.2.
About 60 DAT, paddy canopies cover most of the surface area with EVI values
ranges around 0.6 to 0.7. At the end of the growth period prior to harvesting (the
ripening stage), there is a decrease of EVI values (80 – 105 DAT) around 0.3. In
the fallow period after harvest EVI values is about 0.17 – 0.2.
Table 1 EVI distribution of paddy
No
DAT
1
0–5
2
5 – 10
3
10 – 15
4
15 – 20
5
20 – 25
6
25 – 30
7
30 – 35
8
35 – 40
9
40 – 45
10
45 – 50
11
50 – 55
12
55 – 60
13
60 – 65
14
65 – 70
15
70 – 75
16
75 – 80
17
80 – 85
18
85 – 90
19
90 – 95
20
95 – 100
21
100 – 105
22
End
Source: Domiri et al. (2005)
EVI
< 0 – 0.102
0.103 – 0.139
0.140 – 0.192
0.211 – 0.255
0.256 – 0.327
0.328 – 0.402
0.403 – 0.478
0.479 – 0.551
0.552 – 0.617
0.618 – 0.672
0.673 – 0.714
0.715 – 0.739
0.682 – 0.738
0.637 – 0.681
0.580 – 0.636
0.517 – 0.579
0.450 – 0.516
0.386 – 0.449
0.327 – 0.385
0.278 – 0.326
0.243 – 0.277
0,193 – 0.211
Paddy fields have a physical unique feature that paddy plants are grown on
flooded soils. During the flooding and transplanting period, the land surface is a
mixture of surface water and green paddy plants. Water depth generally varies
from 2 to 15 cm. There is a decrease of leaf and stem moisture content and a
decrease of the number of green leaves at the end of the growth period prior to
harvesting (Xiao et al. 2002). This stage also called fallowing wet. The green
7
color that produced by paddy leaves were at the panicle and called by heading
stage. This stage will be detected by the EVI value as the highest value or in the
other word the crop has the greenest color. The EVI value will decrease when the
paddy started to flowering and then harvested.
Weed Control in Precision Farming
Environmental condition determines the type of equipment and methods of
weed control. It is include a daily weather air temperature, air humidity, wind
speed, and precipitation. It became more complex with other factors such as the
type and age of the main crops, the type and density of weeds in the fields,
availability of herbicides, and the time available for weed control activities.
Seminar et al. (2006) stated that the complexity of these problems can be
overcome if there is a system that able to give a decision which type of equipment
and methods to be used, and it is able to control field activities involving multiple
agents work cooperatively and collaboratively.
Multi-agent system is a method of handling an issue that supported by
proactive agents which able to read changes in environmental conditions and
provide appropriate treatment changes rapidly. While the supervisory control
system serves as a hub that has the ability to choose the type of equipment,
organize, coordinate and integrate the existing units in the system (Solahudin
2013). Astika et al. (2011) designed a real-time data acquisition system using
digital camera and image processing for mapping the closure rate and diversity of
weeds on the field. Then, the map was used as guidance in spraying weed using
variable rate application. Rotinsulu (2011) developed a tool that can detect and
determine weed density of an image and developed a sprayer pump controller
system. Those previous researches have focused on how to distinguish weeds with
plants, then mapping the density of weeds. The map is used as a guidance of
spraying activities in order to realize precision agriculture in terms of provision of
herbicides at the proper location. It is more effective and efficient than the
conventional way.
Relationship with this research is improvement in achieving precision
activities. Research that conducted Astika et al. (2011) is a sub-system of weed
control management system that focuses on the factors and weed plants. By
combined with this research which focuses on weather condition, weed control
activities become right place, right method and right time. Solahudin (2013)
developed a supervisory control system that functioned as a central which ability
to decide the type of tool, managing, coordinating, and integrating every element
of system. This system consist of agents that directly work together read the
change of environmental condition and give a treatment based on the change itself.
Developed supervisory system consists of input module, control knowledge,
climate knowledge, I/O knowledge, crop knowledge and supervisory control
engine. In operational phase supervisory control engine will connect user
preference through input module and knowledgebase to give a task to agents (see
Figure 2). The agents are functioned for: image acquisition, filtering, crop
detection, determination weed density, and determination herbicide dosage.
8
Figure 2 Architecture of supervisory control system based on
precision farming (Solahudin 2013).
That supervisory system has two main functions. The first is consultation
system (off-farm) and the second is spraying system controller (on-farm).
Supervisory system works with all the existing entries in the object processed by
the inference engine based on knowledge modules to determine the best method
for precision weed control activities. Weather-related objects will be developed
further with this research. We try to improve the supervisory system in terms of
its ability to provide consulting based on weather conditions.
3 METHOD
Study Location
Location of study is paddy fields located in Jonggol, Bogor district, West
Java, Indonesia (upper left corner: 6o25‟S 106o7‟E; lower right corner 6o36‟S
107o8‟E) and has an area of 135.65 km2 (Figure 3). That location selected because
there are many paddy field that relative wide. It is enables for homogeneity
example acquisition in a pixel of MODIS image. About 64.3% of Jonggol is
agricultural area, with land use as follows: paddy field, mixed gardens, and
plantations. Paddy field covers about 51.3 km2 or 37.8% of the total area.
According BPS (2008), Jonggol is the largest producer of paddy every year in
Bogor, it is often referred to as the central of rice in Bogor district.
9
LANDUSE
Forest
Plantations
Settlement
Water body
Mixed Garden
Bareland
Rice field
Figure 3 Location of study, Jonggol, Bogor district, West Java, Indonesia.
We determined several paddy fields for samples which have area ≥ 500 m2.
It was related with the highest spatial resolution of each satellite data. These fields
were presented by several pixels of MODIS image while all pixels considered as
one grid coverage of NOAA and TRMM data. These pixels represented paddy
field in Jonggol. To convince pixels as paddy field, all pixels of image were
stacked with landuse map (Figure 4), and used Bing maps for visual interpretation
(Figure 5).
Figure 4 Representation of paddy field area in MODIS
Image (One pixel = 500 x 500 m).
10
Figure 5 Examples of paddy field area in one pixel 500 x 500 m
(www.bing.com/maps).
Time of Research
This research was done from July until December 2013. It conducted in
Remote Sensing/GIS Laboratory of Master of Science in IT for Natural Resources
Management, Bogor Agricultural University, SEAMEO BIOTROP.
Figure 6 General framework of study.
11
Data
Several datasets were used in this research show in Table 2. The main type
of satellite dataset and source described as follow:
MODIS EVI
The Moderate Resolution Imaging Spectroradiometer (MODIS) product
used in this study is the Vegetation Indices (VI) Composite 16-day Global 500 m
SIN Grid V005 or MOD13A1 product, which provided the needed vegetation
phenology data. In addition, the product had already been systematically corrected
for the effects of gaseous and aerosol scattering. The MODIS EVI is embedded in
the MOD13A1 product. The MODIS Land Discipline Group (MODLAND 2010)
developed the EVI for use with MODIS data following this equation:
EV =
where,
,
, and
are the remote sensing reflectances in the nearinfrared, red and blue, respectively, L is a soil adjustment factor and C1 and C2
describe the use of the blue band in correction of the red band for atmospheric
aerosol scattering. The coefficients, C1, C2 and L, are empirically determined as
6.0, 7.5 and 1.0, respectively. G is a gain factor set to 2.5 (Huete et al. 1997;
Setiawan et al. 2011). The EVI data were developed in the above equation in
order to optimize the vegetation signal with improved sensitivity in high biomass
regions. The EV also minimizes atmospheric influences with the „aerosol
resistance‟ term which uses the blue band to correct aerosols influence in band red
(Huete et al. 1997).
In this study we used the MODIS EVI data sets which were acquired from
January 2010 to December 2012 and captured 69 time series with the interval time
16 days. The study area is covered by only one MODIS tile: h28v09. MODIS EVI
data were extracted from the MODIS VI product (MOD13A1) using the MODIS
Reprojection Tool (USGS LP DAAC 2009b) and the selected output format was
GeoTIFF and coordinate system was geographic coordinate systems on datum
World Geodetic System of 1984.
NOAA NCEP Reanalysis 1
Weather data obtained from National Oceanic & Atmospheric
Administration (NOAA) that issued by The Mission of the ESRL Physical
Sciences Division (PSD). The NCEP Reanalysis 1 project is using a state of the
art analysis/forecast system to perform data assimilation using past data from
1948 to the present. It has temporal resolution of 4-times daily, daily and monthly
values for 1948/01/01 to present which has grid global of spatial resolution.
Weather data used in this study were as follows: air temperature, relative humidity,
u-wind and v-wind. Each variable has near the surface (.sig 995 level) dataset on a
2.5ᵒ ×2.5ᵒ grid in daily resolution. The product (.sig 995 level), air temperature,
relative humidity and wind are above surface exactly 2 m, 2 m and 10 m,
12
respectively. For this study we used four kinds of NCEP Reanalysis 1 of data sets
which were acquired from January 2003 to December 2012 and collected 3650
time series for each parameter with daily interval time.
TRMM 3B42
The rainfall product from Tropical Rainfall Measuring Mission (TRMM)
satellite is combination of the Precipitation Radar (PR), TRMM Microwave
Image (TMI), and Visible and Infrared Scanner (VIRS) (Huffman et al. 2007).
TRMM 3B42 daily data is the data level 3 the results of data processing 1B01,
2A12, 3B31, 3A44 and Global precipitation index (GPI). The final 3B42
precipitation (in mm/h) estimates have a 3-hourly temporal resolution and a 0.25ᵒ
x 0.25ᵒ spatial resolution. Spatial coverage extends from 50 degrees south to 50
degrees north latitude. The daily accumulated (beginning at 00Z and ending at
21Z; unit: mm) rainfall product is derived from this 3-hourly product. The data are
stored in flat binary. In this study we used this product which were acquired from
January 2003 to December 2012 and collected 3650 time series with daily interval
time.
Table 2 List of datasets of research
No
1.
2.
3.
4.
5.
Dataset
Topographical map scale 1:250000
Satellite Image 50 x 50 m
MOD S EV , 500 x 500 m
Data from 2010 – 2012
Daily weather data, 2.5 x 2.5 deg
Data from 2003 – 2012
- Temperature
- Relative Humidity
- U-wind and V-wind
Daily Rainfall data, 0. 25 x 0.25 deg
Data from 2003 – 2012
Source
eospatial nformation Agency
www.bing.com/maps
Terra MOD S
NOAA NCEP Reanalysis 1
TRMM
Tools Requirement
Several software and hardware were used during this research such as
personal computer, printer, Remote Sensing and GIS software (ArcGIS 9.3, ENVI
4.7, MODIS Tool), Grads v2.0, climate data operator (cdo), Bing Maps and
Visual Basic 6 programming language.
Data Processing
MODIS EVI data obtained in GeoTIFF format. EVI was extracted using
MODIS conversion toolkit or MODIS re-projection tool that provided by USGS.
MODIS has systematically corrected but not geometrically corrected so that
13
necessary rectified manually. The rectification was done use the corrected beach
vector (Rusdiyatmoko and Zubaidah 2005).
Weather data obtained in netCDF format and geometrically corrected. The
cdo used to manipulate netCDF data format. For example, it used to compute
wind speed and direction which derived from northern and southern wind. The
next step was layer stacking. Both MODIS and weather data were sequentially
stacked to produce the time-series data set and then clipped to cover study area
composite period. The stacked data were evaluated to get temporal pattern from
the time series data (Sakamoto et al. 2005).
Data Analysis
Several points of study area were taken that represented location of paddy
fields. The EVI of these points were analyzed time-series every 16 day during
three years. Weather condition of a pixel weather data where points located was
considered as weather condition of all point of study area because it has coarse
spatial resolution. Weather condition was analyzed time-series daily during 10
years observation. The years observation of EVI data was less than weather data
because the limitation of data storage and for weather study we need long timeseries data.
Identification of Paddy Phenology
We used EVI in our study to detect the phenological stages of paddy. In
order to determine weed control time using spray application, we identified
planting date, heading date, and harvesting date of paddy trough EVI.
Heading time
Heading date used to identify the phenological stages of paddy. According
to time-series data of the spectral reflectance of paddy fields (Domiri et al. 2005),
the maximum EVI appears around the heading date, 60 DAT (Sari et al. 2010;
Semedi 2012). The paddy changes its growth stage from vegetative growth to
reproductive growth on reaching the heading date, and leaves begin to wither and
die. Therefore, we defined the date of the maximum EVI in the time profile as the
estimated heading date and EVI > 0.5 as the threshold.
Planting time
In general, paddy fields are plowed and flooded before planting. The EVI of
paddy fields decreases during this period and then increases again after planting.
Harvesting time
After the heading season, values of EVI begin to decrease as leaves wither
and die and the water in paddy field is dried off before harvesting. EVI then
14
decreases abruptly because of harvesting (Sakamoto et al. 2005). EVI value at the
harvesting stage is around 0.24 – 0.28 (Domiri et al. 2005).
Estimation of Weed Control Time
Normally, weed control in precision agriculture is performed twice, i.e. preplanting and post-emergence (Solahudin 2013). That time estimated after the
planting time known. As mentioned in previous section, paddy phenology from
planting to harvesting represented trough EVI, and the planting time used as
reference to estimate time for weed control. Daily weather condition during preplanting to post-emergence interval was analyzed. Pre-planting estimated a month
before planting month and post-emergence estimated a month after planting
month. Then the interval from pre-harvest to post-emergence considered as weed
control time.
Weed control in paddy cultivation can be performs in various ways. In
Indonesia, generally farmer conducts soil tillage by plowing paddy field before
planting. It purposed to make soil suitable for planting paddy by mixing the
remaining paddy biomass with soil. The paddy field became fertile. It also can kill
the weed. Soil tillage is the first step on weed control in paddy cultivation. Spray
herbicide performed after planting. Time after planting until post-emergence and
it consider as time for spraying herbicide. It is considered the critical period of
weed competition about 40 days after planting.
Development Application to Determine Nozzle Sprayer
This application was developed with objective to determine the proper size
of nozzle sprayer. It expected could minimize drift on weed control. This simple
application could be combined with agents of weed control system (Solahudin
2013) to improve precision of spray application. The rules (IF – THEN) to decide
which nozzle sprayer were acquired from previous researches (Jason 2009;
Solahudin 2013; Tepper 2012) which interpreted into decision tree.
The application was designed and developed using system development life
cycle (SDLC) (O‟Brien 1999). System Development Life Cycle (SDLC) is the
traditional methodology used to develop, maintain, and replace information
system (Hoffer et al. 2007). The phases of the SDLC are: investigation, analysis,
design, implementation and maintenance.
Some rules of decision making based on weather conditions were
(Solahudin 2013):
1) Rules for parameter Wind Velocity (km/h)
- If WV < 2, then do not spray
f 2 ≤ WV ≤ 3, then spraying with a medium droplet and air assist
f 4 ≤ WV ≤ 6, then use a fine droplets
f 7 ≤ WV ≤ 10, then use a medium droplet
f 11 ≤ WV ≤ 14, then use a coarse droplet
f 15 ≤ WV ≤ 20, then spraying with c
VARIABILITY TO SUPPORT DECISION MAKING
ON WEED CONTROL
RIZKY MULYA SAMPURNO
GRADUATE SCHOOL
BOGOR AGRICULTURAL UNIVERSITY
BOGOR
2014
STATEMENT
I, Rizky Mulya Sampurno, hereby declare that this thesis entitled
Characterizing Temporal Dynamic of Weather Variability to Support Decision
Making on Weed Control
Is a result of my work under the supervision advisory board and that it has not been
published before. The content of the thesis has been examined by the advisory board
and external examiner.
Bogor, February 2014
Rizky Mulya Sampurno
G051110091
RINGKASAN
RIZKY MULYA SAMPURNO. Karakterisasi Dinamika Temporal Variabel
Cuaca untuk Mendukung Pengambilan Keputusan pada Pengendalian Gulma.
Dibimbing oleh KUDANG BORO SEMINAR dan YULI SUHARNOTO.
Pada pertanian modern, herbisida dominan digunakan untuk pengendalian
gulma. Meskipun herbisida memiliki manfaat positif dalam memberantas gulma,
herbisida juga berpotensi dapat mencemari lingkungan apabila terjadi spray drift
pada saat penyemprotan. Spray drift dapat terjadi karena kondisi cuaca yang tidak
sesuai. Pengetahuan tentang kondisi cuaca akan membantu petani dan pengambil
keputusan untuk menentukan teknologi dan metode tepat guna dalam
memberantas gulma serta meminimalkan spray drift dan pemborosan herbisida
lainnya.
Kemajuan teknologi informasi telah diterapkan secara luas di bidang
pertanian seperti halnya pertanian presisi. Solahudin (2013) mengembangkan
metode pengendalian gulma berbasis multi-agen. Sistem tersebut memiliki dua
fungsi yaitu fungsi konsultasi dengan memberikan rekomendasi sebelum
melakukan penyemprotan (off-farm) dan pengendalian selama aplikasi
penyemprotan (on-farm). Sistem ini telah diterapkan dalam pengendalian gulma
tanaman kacang tanah.
Hubungan dengan penelitian ini adalah sistem yang dibangun oleh
Solahudin (2013) perlu ditingkatkan kemampuannya dalam hal yang berhubungan
dengan data cuaca. Data cuaca dapat digunakan untuk mengoptimalkan jadwal
penyemprotan pada kegiatan pengendalian gulma yang ramah lingkungan serta
untuk menyiapkan peralatan dan mesin yang akan digunakan sebelum hari
penyemprotan. Kondisi cuaca baik spasial maupun temporal merupakan informasi
yang penting bagi kegiatan pertanian. Integrasi satelit meteorologi dengan produk
prediksi cuaca numeric cukup menjanjikan untuk digunakan sebagai masukan
pengambilan keputusan dalam aplikasi penyemprotan terutama untuk daerah yang
kurang terjangkau oleh stasiun cuaca.
Tujuan dari penelitian ini adalah untuk mengembangkan sistem pendukung
keputusan (DSS) untuk penjadwalan pengendalian gulma dan untuk memilih
ukuran nozzle semprot yang dapat meminimalkan dampak negatif terhadap
lingkungan. Data utama yang diperlukan untuk sistem yang diusulkan ini meliputi
data cuaca 10 tahun yang diperoleh dari satelit penginderaan jauh seperti Oceanic
& Atmospheric Administration Nasional (NOAA), Tropical Rainfall Measuring
Mission (TRMM) dan data indeks vegetasi dari MODerate Resolution Imaging
Spectroradiometer (MODIS). Data indeks vegetasi digunakan untuk menentukan
periode tanam padi. Sedangkan data cuaca digunakan untuk menentukan jadwal
penyemprotan dan untuk menentukan ukuran nozzle semprot yang tepat
berdasarkan kondisi cuaca.
Dalam pertanian presisi, pengendalian gulma dilakukan dua kali yaitu preplanting dan post-emergence. Untuk mengetahui periode ini, kami menggunakan
waktu tanam sebagai referensi. Kami menggunakan tanaman padi untuk penelitian
ini. Waktu tanam tanaman padi mudah diidentifikasi melalui analisis multi
temporal indeks vegetasi yang diperoleh dari penginderaan jauh. Indeks vegetasi
EVI padi pada areal sawah di daerah penelitian dapat menunjukkan siklus
pertumbuhan padi tahunan. Berdasarkan analisis dari EVI, diperkirakan bahwa
waktu penyemprotan padi adalah pada bulan April-Mei untuk siklus tanam 1dan
Oktober-November untuk siklus tanam 2.
Karakteristik pola cuaca di Jonggol diamati selama sepuluh tahun. Setiap
parameter cuaca memiliki karakteristik tersendiri dan umumnya memiliki pola
yang sama dar tahun ke tahun. Pada umumnya, selama sepuluh tahun curah hujan
yang tinggi terjadi pada akhir tahun sampai dengan awal tahun. Sedangkan curah
hujan rendah pada pertengahan tahun. Kecepatan angin berfluktuasi. Angin yang
tinggi terjadi pada akhir tahun sampai Januari dengan kecepatan lebih dari 10
km/jam. Selama sepuluh tahun, suhu minimum dan suhu maksimum adalah
23.57°C dan 30.5°C. Kelembaban relatif menurun bersamaan dengan kenaikan
suhu udara. Kelembaban berkisar antara 67.5 – 95.5%. Kelembaban tinggi terjadi
pada akhir tahun sampai awal tahun, dan kelembaban relatif menurun pada
pertengahan tahun. Petani dan pengambil keputusan dapat menggunakan
informasi cuaca tahun sebelumnya untuk mengetahui waktu yang optimal untuk
penjadwalan aplikasi penyemprotan, untuk mempersiapkan alat dan mesin
termasuk sprayer. Waktu yang optimal untuk pengendalian gulma ditentukan dari
interval waktu penyemprotan yang diperoleh dari analisis indeks vegetasi dengan
mempertimbangkan kondisi cuaca.
Aplikasi untuk menentukan ukuran nozzle semprot telah dikembangkan
berdasarkan kriteria-kriteria kondisi cuaca. Pengetahuan untuk menentukan
ukuran nozzle diperoleh dari penelitian sebelumnya. Curah hujan dijadikan
sebagai parameter utama dalam menentukan keputusan untuk melakukan
penyemprotan. Parameter selanjutnya adalah angin, suhu, dan kelembaban.
Parameter cuaca dapat diinput secara manual atau dapat menggunakan data cuaca
yang tersimpan database.
Sistem penunjang keputusan ini telah dikembangkan dan diuji dengan data
riil di lapangan yang diperoleh dari satelit penginderaan jauh. Sistem yang
dikembangkan dapat memberikan
informasi waktu yang optimal untuk
penyemprotan pada tanaman padi serta dapat menentukan ukuran nozzle semprot
berdasarkan kondisi cuaca. Tidak hanya untuk tanaman padi, sistem ini dapat
diterapkan baik pada tanaman di dataran rendah maupun di dataran tinggi.
Kata kunci: DSS, herbisida, spray drift, pola cuaca, pengendalian gulma
SUMMARY
RIZKY MULYA SAMPURNO. Characterizing Temporal Dynamic of Weather
Variability to Support Decision Making on Weed Control. Supervised by
KUDANG BORO SEMINAR and YULI SUHARNOTO.
Herbicide is the dominant tool used for weed control in modern agriculture.
Although herbicide has positive benefit in killing the target weeds, it potential
becomes negative impact to the environment if some remains in the air and drift.
Spray drift can happen due to unsuitable weather. The knowledge of weather
condition will help farmer and decision maker to decide the appropriate
technology and method for eradicating weed which minimize drift and other
potential waste.
The progress of information technology has been applied widely in
agriculture such as precision agriculture. Solahudin (2013) developed a weed
control method using multi-agents based. That system has two functions i.e.
consultation function by giving recomendation before spraying (off-farm) and
controlling during spray application (on-farm) using multi-intelligent agents
which applied for groundnut farming.
Relationship with this research is the system built by Solahudin (2013)
needs to be improved in knowledge that related with weather. Weather data can be
used to optimize spray scheduling on weed control activities that safe to the
environment and for setup the equipment and machinery which prepared earlier
before application day. The spatial and temporal variability weather conditions are
important sources for agricultural activities such as spray application. Integration
meteorological satellite with numerical weather prediction (NWP) product is
promising in find timely weather variables as input for decision making to resolve
problems in spray application especially for area which sparse coverage of
weather stations.
The objective of this research is to develop a decision support system (DSS)
for schedulling of weed control and for selecting the proper nozzle size of the
sprayers that introduce minimum negative impact to the environment. The main
set of data required for our proposed system includes the set of 10 years weather
data series acquired from remote sensing such as the National Oceanic &
Atmospheric Administration (NOAA) and the Tropical Rainfall Measuring
Mission (TRMM) and a set of vegetation index from the MODerate Resolution
Imaging Spectroradiometer (MODIS). The vegetation index data utilized to
determine the planting period of paddy and weather data set utilized to determine
spray schedule and to determine the proper size of the sprayers for weed control.
In precision farming, weed control is done two times i.e. pre-planting and
post-emergence. To know these times, we used planting time as reference. We
used paddy for this study. Paddy planting time is easy to identify through multitemporal analysis of vegetation index. Enhanced vegetation index (EVI) of paddy
field in study area shows the annual paddy growth cycle, it representing intensive
cropping with multiple harvests. Based on analysis of EVI, we estimated the
spraying times for paddy are April to May and October to November for cycle 1
and cycle 2, respectively.
We study weather pattern in Jonggol during ten years. Every parameter have
own characteristic and generally in same fluctuated pattern form. Generally,
during ten years rainfall is high in year end to early year while low in mid-year.
Wind speed is fluctuates. Wind is high in year end to January every year about
more than 10 km/s. For ten years, minimum and maximum temperatures are
23.57°C and 30.5°C. Relative humidity decreased when air temperature increase.
It is about 67.5 – 95.5%, high in year end to early year, and lowers in middle year.
Farmer or decision maker can use information from past weather data to find out
optimal time for scheduling, preparing machinery and sprayer. Optimal week for
weed control determined from interval time for spraying both in crop cycle 1 and
crop cycle 2.
We developed application to determine the proper nozzle size for sprayer
based on weather condition. Knowledge to determine nozzle size is acquired from
previous research. Rainfall is the first parameter which decides do spray or do not
spray, because spray application will not conducted in rainy day and herbicide
particles will run off along with rain water. Wind becomes second parameter,
following by temperature and humidity. Weather parameters can be inputted
manually or can use weather data taken from the past datasets which stored in
database.
The DSS for weed applications has been developed and tested with a real
data set acquired from remote sensing devices. The developed system can
generate optimal spray scheduling and recommend the proper size of nozzles used
for spray application on paddy crops based on the weather condition, and thus
minimizing spray drift and bad environmental impact. This method could be
implemented both for low-land crop and high-land crop.
Key words: DSS, herbicides, spray drift, weather pattern, weed control
Copyright © 2014, Bogor Agricultural University
Copyright are protected by law
It is prohibited to cite all of part of this thesis without referring to and mentioning the
source. Citation only permitted for the sake of education, research, scientific writing,
report writing, critical writing or reviewing scientific problem. Citation does not
inflict the name and honor of Bogor Agricultural University
It is prohibited to republish and reproduce all part of this thesis without written
permission from Bogor Agricultural University
CHARACTERIZING TEMPORAL DYNAMIC OF WEATHER
VARIABILITY TO SUPPORT DECISION MAKING
ON WEED CONTROL
RIZKY MULYA SAMPURNO
A thesis submitted for the degree Master of Science in Information
Technology for Natural Resources Management Study Program
GRADUATE SCHOOL
BOGOR AGRICULTURAL UNIVERSITY
BOGOR
2014
The External Examiner:
Dr Liyantono, STP, MAgr
Thesis Title : Characterizing Temporal Dynamic of Weather Variability to
Support Decision Making on Weed Control
Name
: Rizky Mulya Sampurno
Student ID : G051110091
Approved by
Advisory Board
Prof Dr Ir Kudang Boro Seminar, MSc
Supervisor
Dr Ir Yuli Suharnoto, MEng
Co-Supervisor
Endorsed by
Program Coordinator of MSc in IT
for Natural Resources Management
Dr Ir Hartrisari Hardjomidjojo, DEA
Date of Examination:
January, 29th 2014
Dean of Graduate School
Dr Ir Dahrul Syah, MScAgr
Date of Graduation:
Thesis Title : Characterizing Temporal Dynamic of Weather Variability to
Support Decision Making on Weed Control
:
Rizky Mulya Sampumo
Name
Student ID : GOS111 0091
Approved by
Advisory Board
Prof Dr Ir Kudang Boro Seminar, MSc
Supervisor
Dr Ir Yuli Suhamoto, MEng
Co-Supervisor
Endorsed by
Program Coordinator of MSc in IT
for Natural Resources Management
Dr lr Hartrisari Hardjomidjojo, DEA
Date of Examination:
1h
January, 29 2014
Date of Graduation :
28
MAR 2014
ACKNOWLEDGMENTS
Alhamdulillah, first and most, I thank to The Almighty Allah subhanahu wa
ta’ala for allowing me alive and get many experiences in this world. Plenty rahmats
have been given makes me easier to complete this thesis. The topic selected for this
research is utilization of remote sensing technology for agricultural application that
conducted from July – November 2013 with the title Characterizing temporal
dynamic of weather variability to support decision making on weed control.
I owe a great deal to Prof Dr Kudang Boro Seminar who served as chair of the
board of supervisor and teach anything. Dr Yuli Suharnoto as co-supervisor made
significant contribution to many aspects in the thesis, particularly in remote sensing.
Thank you very much Pak Kudang and Pak Yuli for guidance, patience, comments,
correction and constructive inputs through all months of my research.
I would like to thank Dr Mohamad Solahudin for inputs and idea for this
research. My appreciation also goes to Dr Liyantono and Dr Bib Paruhum Silalahi as
the external examiners for their valuable critics, inputs, and corrections. I would like
to thank Dr Ibnu Sofyan, Mr Harry Imantho, Dr Tofael Ahamed, Dr Ryozo Noguchi,
Dr Yudi Setiawan for the valuable inputs and comments. I would like to thank to all
MIT secretariat that support our administration, technical and facility. I would like to
thank Agus Aris and all my colleagues in MIT for helping, supporting, and
togetherness. Special thanks go to Nadia Tannia Hendartina for any help and support
during my study.
This work is dedicated to my parent, Ibu and Bapak who cheer my life at the
beginning and afterward, prayer, financial, encouragement and support. Hopefully,
the results of this research would provide a positive and valuable contribution for
anyone who reads it.
Bogor, February 2014
Rizky Mulya Sampurno
LIST OF CONTENTS
LIST OF TABLES
vii
LIST OF FIGURES
vii
LIST OF APENDICES
vii
LIST OF ABREVIATON
viii
1 INTRODUCTION
Background
Problem Formulation
Objective
Research Benefits
Scope of Research
1
1
2
2
3
3
2 LITERATURE REVIEW
Weed
Weed Control on Paddy Field using Herbicide
Spray Drift
Paddy Phenological Stages
Weed Control in Precision Farming
3
3
4
5
5
7
3 METHOD
Study Location
Time of Research
Data
Tools Requirement
Data Processing
Data Analysis
Identification of Paddy Phenology
Estimation of Weed Control Time
Development Application to Determine Nozzle Sprayer
8
8
10
11
12
12
13
13
14
14
4 RESULT AND DISCUSSION
Paddy Planting Time
Weed Control and Spray Application Time
Weather Pattern
Application for Minimizing Spray Drift
Herbicide Application in Agricultural crop
15
16
17
18
21
22
5 CONCLUSION AND RECOMMENDATION
Conclusion
Recommendation
23
23
23
REFERENCES
23
APPENDICES
27
CURRICULLUM VITAE
34
LIST OF TABLES
1
2
3
EVI distribution of paddy
List of datasets of research
Decision making to determine nozzle size based on weather parameter
6
12
15
LIST OF FIGURES
1
2
3
4
5
6
7
8
9
10
11
12
Weed on some main crops; (a) paddy, (b) groundnut, and (c) maize
Architecture of supervisory control system based on precision farming
(Solahudin 2013)
Location of study, Jonggol, Bogor district, West Java, Indonesia
Representation of paddy field area in MODIS Image (One pixel = 500
x 500 m)
Examples of paddy field area in one pixel 500 x 500 m
(www.bing.com/maps)
General framework of study
Utilization of remote sensing data for weed control
Weed control time which estimated from planting time through
enhanced vegetation index (EVI)
Weather patterns of study area and it utilization to characterize
weather condition during spray duration
Weather condition during spray application interval in both paddy
planting cycle 1 period March – May (a) and paddy planting cycle 2
period September – November (b)
Knowledge representation of nozzle selection based on weather
condition; P: rainfall, W: wind speed, T: temperature, and RH: relative
humidity
Sprayer nozzle selection based on normal weather condition, (a)
application in normal condition and (b) application in extreme
condition
3
8
9
9
10
11
16
17
18
20
21
22
LIST OF APPENDICES
1
2
3
4
Coordinate of several paddy fields in Jonggol
27
Enhanced vegetation index (EVI) value of several paddy fields in Jonggol 28
Examples of processing netCDF data
29
Sourcecode of Application for Minimizing Spray Drift
31
LIST OF ABBREVIATION
BPS
cdo
DAP
DSS
EVI
ESRL
GeoTIFF
GIS
GPI
I/O
LAPAN
LP DAAC
LSWI
MODIS
NCEP
NDVI
NDWI
netCDF
NOAA
NWP
PR
PSD
RH
SDLC
TMI
TRMM
USGS
VIRS
WV
Badan Pusat Statistik
Climate Data Operator
Days After Planting
Decision Support System
Enhanced Vegetation Index
Earth System Research Laboratory
Geostationary Earth Orbit Tagged Image File Format
Geographic Information System
Global precipitation index
Input/Output
Lembaga Penerbangan dan Antariksa Nasional
Land Processes Distributed Active Archive Center
Land Surface Water Index
MODerate Resolution Imaging Spectroradiometer
National Centers for Environmental Prediction
Normalized Difference Vegetation Index
Normalized Difference Water Index
Network Common Data Form
National Oceanic & Atmospheric Administration
Numerical Weather Prediction
Precipitation Radar
Physical Sciences Division
Relative Humidity
System Development Life Cycle
TRMM Microwave Image
Tropical Rainfall Measuring Mission
United State Geological Survey
Visible and Infrared Scanner
Wind Velocity
1
1 INTRODUCTION
Background
Weeds are a serious problem for agricultural crop. They rob main crops of
sunlight, water and nutrient causing production losses both in quantity and quality
(Solahudin 2013). Losses due to weed are: wheat (9.8%), rice (10.8%), maize
(13%), sorghum (17.8%), potatoes (4%) and groundnut (11.8%) (Evans 1996).
Even, an uncontrolled weed can decrease yield until 20-80% (Utami 2004). In
modern way, agriculture activity is depends on utilization of chemical substance
to increase production. Herbicides are the dominant tool used for weed control in
modern agriculture (Harker et al. 2013).
Although herbicide has positive benefit in killing the target weeds, it
potentially becomes negative impact if some remains in the air and drift. Spray
drift from herbicide can cause crop protection chemicals to be deposited in
undesirable areas (Nuyttens et al. 2011). It has serious consequences such as
damage to sensitive adjoining crops (Reddy et al. 2010), damage to susceptible
off-target areas (Marrs and Frost 1997), environmental contamination, illegal
herbicide residues, lowers yield results (Kjaer et al. 2006) and health risks to
animals (Otto et al. 2009) and people (Butler et al. 2010).
Spray drift continues to be a major problem in applying herbicides. Factors
that cause drift are weather conditions and sprayer setup (Wolf 2006). Drift can
happen due to unsuitable weather. It potentially occurred every time when sprayer
turned on. The knowledge of weather condition will help farmer and decision
maker to decide the appropriate technology and method for eradicating weed, plan,
and effectively execute spray applications to avoid spray drift and other potential
waste.
The progress of information technology has been applied widely in
agriculture such as precision agriculture (Auernhammer 2001). Solahudin (2013)
has developed a weed control method in precision agriculture using multi-agents
based. The method is a supervisory system to determine technology and liquid
applicator capacity and controlling agents. That system has two functions i.e. as
consultation function by giving recomdation before spraying (off-farm) and
controlling during spraying using multi-intelligent agents (on-farm) which applied
for groundnut farming. Decision making method influence by factors on weeding
activity such as time schedule, total area, working time, type of crop, type of weed,
weed density, herbicide, weather, and sprayer technology. While agents are for:
image acquisition, filtering, crop detection, determination weed density, and
determination herbicide dosage.
The weather conditions both spatially and temporally have not studied more
in previous research. Relationship with this research is the system built by
Solahudin (2013) needs to be improved in knowledge that related with weather.
Weather data can be used to optimize spray scheduling in weed control activities
to avoid drift and environmental friendly. Also it can be used in preparing the
equipment and machinery that determine earlier before application day.
The spatial and temporal variability weather conditions are important
sources for agricultural activities such as spray application. Integration
2
meteorological satellite with the Numerical Weather Prediction (NWP) product is
promising in find timely weather variables as input for decision making to resolve
problems in spray application especially for area which sparse coverage of
weather stations (Bahar 2012; Roebeling et al. 2004; De Wit et al. 2010).
However, the availability of data in real-time is still difficult to achieve. The
Tropical Rainfall Measuring Mission (TRMM) data is capable of providing daily
rainfall. NWP products from NCEP/NOAA such as 2 m temperature, wind, and
RH (Relative Humidity) are used as other input. Moreover data from experience
could be used for scheduling and as decision support for preparing tools and
machinery before spray application conducted.
Problem Formulation
In addition to the effectiveness and efficiency issues, drift of chemical
substance on herbicide application can cause damage to other agricultural product
and pollute the environment. Utilization of weather data to resolve these problems
is needed. However, the availability of data in real-time is still difficult to achieve.
Therefore, utilization of remote sensing data is expected can help to resolve those
problems.
Objective
The objective of this research is to develop a decision support system (DSS)
for schedulling of weed spraying and for selecting the proper nozzle size of the
sprayers that introduce minimum negative impact to the environment.
Research Benefits
The benefits of this research are:
- Introduce a method that utilizes weather data for agriculture application.
- Provide information about characteristics of weather conditions on study area.
- Provide recommendation the proper nozzle size of sprayer based on weather
condition.
Scope of Research
This research is limited to the development of weed control methods based
on precision agriculture. The developed method is functioned as decision support
systems for spray scheduling activities and applications to determine the proper
nozzle size of sprayer based on weather conditions. We employ weather and
vegetation data remotely sensed from satellite observation result. It is also use to
overcome the limitation of meteorology station in providing weather data. This
weed control method suitable to be applied on big scale farming area. The main
crop that observed in this study is paddy. Paddy fields easily identified using
3
remote sensing especially for big scale farming area. Paddy becomes main
agricultural product that planted in many region of Indonesia. This developed
DSS can be applied not only for paddy but also for other crop which used spray
application.
2 LITERATURE REVIEW
Weed
Weeds are unwanted plants growing along with agricultural crops (Dangwal
et al. 2011). Weeds compete with crops mainly for space, sunlight, moisture,
nutrient, and reduce the quantity as well as quality of production (Solahudin
2013). Holm et al. (1979) estimated 250 weed species which are harmful for
agricultural crops throughout the world. The competition of weeds for nutrients
may results in such obvious responses as dwarfing in plant size, nutrient starved
conditions, wilting and actual dying out of plants (Anderson et al. 1996). Weed
seeds germinate earlier to agricultural crops, their seedlings grows faster and
aggressive so that they crowd out all other plants which possesses more valuable
properties and establish a kingdom of their own within a short period of time.
Weed species mature ahead of crops so that their seeds are collected with the crop
harvest and get distributed to other places. Some weed species caused damage to
crops by harboring pests and disease agents.
(a)
(b)
(c)
Figure 1 Weed on some main crops; (a) paddy, (b) groundnut, and (c) maize.
Weeds act as host for bacteria, viruses and nematodes that causes diseases in
crop plants (Peters et al. 1955). Weeds show allelopathic effects on agricultural
crops by secreting allelochemicals that inhibit their growth and germination
(Oudhia P and Tripathi RS 1998). The weedy crop may sometime leads to
complete failure. Weeds cause yield losses as reported by Evan (1996) were wheat
(9.8%), rice (10.8%), maize (13%), sorghum (17.8%), potatoes (4%), sugar
(15.7%), soybean (13.5%), cacao (11.9%) and groundnut (11.8%). An
uncontrolled weed can decrease yield until 20-80% (Utami 2004). The cost of
removing weeds adds to the cost of production of crops, thus producers losses part
of their investment and the country suffers a reduction in agricultural products.
4
In irrigated area, competition between weed and paddy can decrease rice
yield 10-40%, depend on species and density of weed, soil type, water and
weather (Nantasomsaran and Moody 1993). Nyarko and De Datta (1991) reported
that yield loss of paddy by weed can be 10-15%, while Pane et al. (2002) in
Karawang, yield loss about 8-12%. It shows that weed control which perform by
farmer is not effective especially for manual weeding. World Bank (1996)
reported in 1995, weed decrease yield in Asia about 50 million tones. Competition
weed and paddy crop decreased yield 10 million tones every year in China
(Labrada 2003).
Weed Control on Paddy Field using Herbicide
There are many methods of weed control. It can be categorized into two
groups, i.e. indirect method and direct method (Pane and Jatmiko 2009). Indirect
methods are: (1) using good seed and variety, (2) soil tillage, (3) irrigation and
fertilizer control, (4) controlling planting space, (5) crop rotation, and (6)
biological method. Direct methods are: (1) hand pulling, (2) mechanic method,
and (3) herbicide.
In high density population country, the salary of worker is relatively
inexpensive. Weed control perform manually by hand pulling and flooding (Pane
and Jatmiko 2009). Recently, hand pulling is not effective but the worker is
limited. Then, weed control is performed using mechanical method. The next
generation is development of chemical method known as herbicide. It is claimed
more efficient, cheap, quick, and minimal worker.
Nowadays, there are many farmers that use herbicide to control weed (Pane
and Jatmiko 2009). In irrigation paddy field such as in outer Java Island, weeding
worker is rare and expensive. In West Java especially irrigated area of Jatiluhur,
because the planting time is same, all farmers need worker while the worker is
limited. Also in paddy field area near the city, the worker is also limited because
younger worker tend to work in office, factory, property etc.
De Datta and Herdt (1983) reported that in agricultural area where the salary
of worker is expensive and well irrigation management area such as Japan, Korea,
and Taiwan, about 75-100% of farmer have used herbicide. According to (Chisaka
1995), combination of herbicide with indirect methods like soil tillage and
fertilization prove more effective than hand pulling. Direktorat Jendral Pertanian
Tanaman Pangan (1982) reported that farmers in Deli Serdang (North Sumatera),
Musi Banyuasin (South Sumatera), Sidrap (South Celebes), and Karawang and
Indramayu (West Java) used herbicide 21%, 37.5%, 100%, and 17.5%,
respectively.
Each plant species has a different critical period in its competition with
weeds. Generally critical period of weed competition starts since the plant grows
to 1/4-1/3 the first of the plant life cycle (Pane and Jatmiko 2009). For example in
paddy, critical period of weed occurs by age the first 40 days of its life cycle. In
this stage, canopy of paddy is not closed and the sunlight intensity is high in the
ground. Weeds seed germinate and grow faster than paddy.
5
Spray Drift
Spray drift is becoming an increasingly important part of every spraying
operation. Spray drift is the movement of a pesticide/herbicide through the air,
during or after application, to a site other than the intended target. Drift is
considered to be the most challenging problem facing applicators and
pesticide/herbicide manufacturers. Although drift may occur as vaporized active
chemical substances from the application site, it is usually the physical movement
of very small drops from the target area at the time of application (Hofman and
Elton 2001).
Spray drift and from agricultural pesticides/herbicides can cause crop
protection chemicals to be deposited in undesirable areas (Nuyttens et al. 2011).
This can have serious consequences such as damage to sensitive adjoining crops
(Reddy et al. 2010), damage to susceptible off-target areas (Marrs and Frost
1997), environmental contamination, illegal pesticide/herbicide residues, lower
yield results (Kjaer et al. 2006) and health risks to animals (Otto et al. 2009) and
people (Butler et al. 2010).
There are several factors that play a significant part in the creation and
reduction of drift (Anonymous 2010; Hofman and Elton 2001). They can be
grouped into one of the following categories: (a) spray characteristics, such as
volatility and viscosity of the pesticide formulation, (b) equipment and application
techniques such as nozzle sprayer and distance between target weed with nozzle
sprayer, (c) weather conditions at the time of application (wind speed and
direction, temperature, relative humidity and stability of air at the application
site), (d) operator care, attitude and skill.
Paddy Phenological Stages
The phenology of paddy is divided into several stages (Semedi, 2012). The
first stage is flooded. It performs before the paddy seed being planted. Once the
seed planted or at germination stage, it will take 25-30 days to reach tillering and
the paddy leaf starting to grow. The leaf color then turning from yellow into green
in tillering stages. The growth will reach maximum at the panicle where for early
variety of paddy it will take 55-60 days from germination stage and for late
variety it will take 65-75 days. It is called heading time. The next stage is the
flowering, paddy started to grow flowers and the leaf color started to turn into
yellow. For the early variety of paddy, this stage reaches in 85-100 days after
germination and for the late variety of paddy it needs 100-115 days. The last stage
is harvesting, where paddy is already developed and ready to be harvested. The
early variety of paddy needs 130-145 days after germination to be harvested,
while for the late variety needs 140-165 days. The paddy phenological stages can
be studied using vegetation index.
A number of researches has been attempted to detect paddy phenological
stages in wide extent of Asian region using MODIS data (Xiao et al. 2005; 2006,
Sakamoto et al. 2005, Uchida 2007, Sari et al. 2010 and Setiawan et al. 2011.
They employed a kind of indices derived from MODIS data such as Enhanced
Vegetation Index (EVI), Normalized Difference Vegetation Index (NDVI), LSWI,
6
and NDWI. Sakamoto et al. (2005) and Uchida (2007) used time-series EVI data
to determine the planting date, heading date, harvesting date, and growing period
of paddy. Sari et al. (2010) study paddy phenology in paddy fields with complex
cropping pattern.
Sari et al. (2010) utilized MODIS EVI to detect growing stages of paddy
crop cycle. Sari et al. (2010) characterized the phenological stages of paddy into
three main periods: (1) the flooding and transplanting period; (2) the growing
period (vegetative growth, reproductive and ripening stages); and (3) the fallow
period after harvest. Acording to Domiri et al. (2005), during flooding and paddy
transplanting period, about 0-20 days after transplanting (DAT), EVI values < 0.2.
About 60 DAT, paddy canopies cover most of the surface area with EVI values
ranges around 0.6 to 0.7. At the end of the growth period prior to harvesting (the
ripening stage), there is a decrease of EVI values (80 – 105 DAT) around 0.3. In
the fallow period after harvest EVI values is about 0.17 – 0.2.
Table 1 EVI distribution of paddy
No
DAT
1
0–5
2
5 – 10
3
10 – 15
4
15 – 20
5
20 – 25
6
25 – 30
7
30 – 35
8
35 – 40
9
40 – 45
10
45 – 50
11
50 – 55
12
55 – 60
13
60 – 65
14
65 – 70
15
70 – 75
16
75 – 80
17
80 – 85
18
85 – 90
19
90 – 95
20
95 – 100
21
100 – 105
22
End
Source: Domiri et al. (2005)
EVI
< 0 – 0.102
0.103 – 0.139
0.140 – 0.192
0.211 – 0.255
0.256 – 0.327
0.328 – 0.402
0.403 – 0.478
0.479 – 0.551
0.552 – 0.617
0.618 – 0.672
0.673 – 0.714
0.715 – 0.739
0.682 – 0.738
0.637 – 0.681
0.580 – 0.636
0.517 – 0.579
0.450 – 0.516
0.386 – 0.449
0.327 – 0.385
0.278 – 0.326
0.243 – 0.277
0,193 – 0.211
Paddy fields have a physical unique feature that paddy plants are grown on
flooded soils. During the flooding and transplanting period, the land surface is a
mixture of surface water and green paddy plants. Water depth generally varies
from 2 to 15 cm. There is a decrease of leaf and stem moisture content and a
decrease of the number of green leaves at the end of the growth period prior to
harvesting (Xiao et al. 2002). This stage also called fallowing wet. The green
7
color that produced by paddy leaves were at the panicle and called by heading
stage. This stage will be detected by the EVI value as the highest value or in the
other word the crop has the greenest color. The EVI value will decrease when the
paddy started to flowering and then harvested.
Weed Control in Precision Farming
Environmental condition determines the type of equipment and methods of
weed control. It is include a daily weather air temperature, air humidity, wind
speed, and precipitation. It became more complex with other factors such as the
type and age of the main crops, the type and density of weeds in the fields,
availability of herbicides, and the time available for weed control activities.
Seminar et al. (2006) stated that the complexity of these problems can be
overcome if there is a system that able to give a decision which type of equipment
and methods to be used, and it is able to control field activities involving multiple
agents work cooperatively and collaboratively.
Multi-agent system is a method of handling an issue that supported by
proactive agents which able to read changes in environmental conditions and
provide appropriate treatment changes rapidly. While the supervisory control
system serves as a hub that has the ability to choose the type of equipment,
organize, coordinate and integrate the existing units in the system (Solahudin
2013). Astika et al. (2011) designed a real-time data acquisition system using
digital camera and image processing for mapping the closure rate and diversity of
weeds on the field. Then, the map was used as guidance in spraying weed using
variable rate application. Rotinsulu (2011) developed a tool that can detect and
determine weed density of an image and developed a sprayer pump controller
system. Those previous researches have focused on how to distinguish weeds with
plants, then mapping the density of weeds. The map is used as a guidance of
spraying activities in order to realize precision agriculture in terms of provision of
herbicides at the proper location. It is more effective and efficient than the
conventional way.
Relationship with this research is improvement in achieving precision
activities. Research that conducted Astika et al. (2011) is a sub-system of weed
control management system that focuses on the factors and weed plants. By
combined with this research which focuses on weather condition, weed control
activities become right place, right method and right time. Solahudin (2013)
developed a supervisory control system that functioned as a central which ability
to decide the type of tool, managing, coordinating, and integrating every element
of system. This system consist of agents that directly work together read the
change of environmental condition and give a treatment based on the change itself.
Developed supervisory system consists of input module, control knowledge,
climate knowledge, I/O knowledge, crop knowledge and supervisory control
engine. In operational phase supervisory control engine will connect user
preference through input module and knowledgebase to give a task to agents (see
Figure 2). The agents are functioned for: image acquisition, filtering, crop
detection, determination weed density, and determination herbicide dosage.
8
Figure 2 Architecture of supervisory control system based on
precision farming (Solahudin 2013).
That supervisory system has two main functions. The first is consultation
system (off-farm) and the second is spraying system controller (on-farm).
Supervisory system works with all the existing entries in the object processed by
the inference engine based on knowledge modules to determine the best method
for precision weed control activities. Weather-related objects will be developed
further with this research. We try to improve the supervisory system in terms of
its ability to provide consulting based on weather conditions.
3 METHOD
Study Location
Location of study is paddy fields located in Jonggol, Bogor district, West
Java, Indonesia (upper left corner: 6o25‟S 106o7‟E; lower right corner 6o36‟S
107o8‟E) and has an area of 135.65 km2 (Figure 3). That location selected because
there are many paddy field that relative wide. It is enables for homogeneity
example acquisition in a pixel of MODIS image. About 64.3% of Jonggol is
agricultural area, with land use as follows: paddy field, mixed gardens, and
plantations. Paddy field covers about 51.3 km2 or 37.8% of the total area.
According BPS (2008), Jonggol is the largest producer of paddy every year in
Bogor, it is often referred to as the central of rice in Bogor district.
9
LANDUSE
Forest
Plantations
Settlement
Water body
Mixed Garden
Bareland
Rice field
Figure 3 Location of study, Jonggol, Bogor district, West Java, Indonesia.
We determined several paddy fields for samples which have area ≥ 500 m2.
It was related with the highest spatial resolution of each satellite data. These fields
were presented by several pixels of MODIS image while all pixels considered as
one grid coverage of NOAA and TRMM data. These pixels represented paddy
field in Jonggol. To convince pixels as paddy field, all pixels of image were
stacked with landuse map (Figure 4), and used Bing maps for visual interpretation
(Figure 5).
Figure 4 Representation of paddy field area in MODIS
Image (One pixel = 500 x 500 m).
10
Figure 5 Examples of paddy field area in one pixel 500 x 500 m
(www.bing.com/maps).
Time of Research
This research was done from July until December 2013. It conducted in
Remote Sensing/GIS Laboratory of Master of Science in IT for Natural Resources
Management, Bogor Agricultural University, SEAMEO BIOTROP.
Figure 6 General framework of study.
11
Data
Several datasets were used in this research show in Table 2. The main type
of satellite dataset and source described as follow:
MODIS EVI
The Moderate Resolution Imaging Spectroradiometer (MODIS) product
used in this study is the Vegetation Indices (VI) Composite 16-day Global 500 m
SIN Grid V005 or MOD13A1 product, which provided the needed vegetation
phenology data. In addition, the product had already been systematically corrected
for the effects of gaseous and aerosol scattering. The MODIS EVI is embedded in
the MOD13A1 product. The MODIS Land Discipline Group (MODLAND 2010)
developed the EVI for use with MODIS data following this equation:
EV =
where,
,
, and
are the remote sensing reflectances in the nearinfrared, red and blue, respectively, L is a soil adjustment factor and C1 and C2
describe the use of the blue band in correction of the red band for atmospheric
aerosol scattering. The coefficients, C1, C2 and L, are empirically determined as
6.0, 7.5 and 1.0, respectively. G is a gain factor set to 2.5 (Huete et al. 1997;
Setiawan et al. 2011). The EVI data were developed in the above equation in
order to optimize the vegetation signal with improved sensitivity in high biomass
regions. The EV also minimizes atmospheric influences with the „aerosol
resistance‟ term which uses the blue band to correct aerosols influence in band red
(Huete et al. 1997).
In this study we used the MODIS EVI data sets which were acquired from
January 2010 to December 2012 and captured 69 time series with the interval time
16 days. The study area is covered by only one MODIS tile: h28v09. MODIS EVI
data were extracted from the MODIS VI product (MOD13A1) using the MODIS
Reprojection Tool (USGS LP DAAC 2009b) and the selected output format was
GeoTIFF and coordinate system was geographic coordinate systems on datum
World Geodetic System of 1984.
NOAA NCEP Reanalysis 1
Weather data obtained from National Oceanic & Atmospheric
Administration (NOAA) that issued by The Mission of the ESRL Physical
Sciences Division (PSD). The NCEP Reanalysis 1 project is using a state of the
art analysis/forecast system to perform data assimilation using past data from
1948 to the present. It has temporal resolution of 4-times daily, daily and monthly
values for 1948/01/01 to present which has grid global of spatial resolution.
Weather data used in this study were as follows: air temperature, relative humidity,
u-wind and v-wind. Each variable has near the surface (.sig 995 level) dataset on a
2.5ᵒ ×2.5ᵒ grid in daily resolution. The product (.sig 995 level), air temperature,
relative humidity and wind are above surface exactly 2 m, 2 m and 10 m,
12
respectively. For this study we used four kinds of NCEP Reanalysis 1 of data sets
which were acquired from January 2003 to December 2012 and collected 3650
time series for each parameter with daily interval time.
TRMM 3B42
The rainfall product from Tropical Rainfall Measuring Mission (TRMM)
satellite is combination of the Precipitation Radar (PR), TRMM Microwave
Image (TMI), and Visible and Infrared Scanner (VIRS) (Huffman et al. 2007).
TRMM 3B42 daily data is the data level 3 the results of data processing 1B01,
2A12, 3B31, 3A44 and Global precipitation index (GPI). The final 3B42
precipitation (in mm/h) estimates have a 3-hourly temporal resolution and a 0.25ᵒ
x 0.25ᵒ spatial resolution. Spatial coverage extends from 50 degrees south to 50
degrees north latitude. The daily accumulated (beginning at 00Z and ending at
21Z; unit: mm) rainfall product is derived from this 3-hourly product. The data are
stored in flat binary. In this study we used this product which were acquired from
January 2003 to December 2012 and collected 3650 time series with daily interval
time.
Table 2 List of datasets of research
No
1.
2.
3.
4.
5.
Dataset
Topographical map scale 1:250000
Satellite Image 50 x 50 m
MOD S EV , 500 x 500 m
Data from 2010 – 2012
Daily weather data, 2.5 x 2.5 deg
Data from 2003 – 2012
- Temperature
- Relative Humidity
- U-wind and V-wind
Daily Rainfall data, 0. 25 x 0.25 deg
Data from 2003 – 2012
Source
eospatial nformation Agency
www.bing.com/maps
Terra MOD S
NOAA NCEP Reanalysis 1
TRMM
Tools Requirement
Several software and hardware were used during this research such as
personal computer, printer, Remote Sensing and GIS software (ArcGIS 9.3, ENVI
4.7, MODIS Tool), Grads v2.0, climate data operator (cdo), Bing Maps and
Visual Basic 6 programming language.
Data Processing
MODIS EVI data obtained in GeoTIFF format. EVI was extracted using
MODIS conversion toolkit or MODIS re-projection tool that provided by USGS.
MODIS has systematically corrected but not geometrically corrected so that
13
necessary rectified manually. The rectification was done use the corrected beach
vector (Rusdiyatmoko and Zubaidah 2005).
Weather data obtained in netCDF format and geometrically corrected. The
cdo used to manipulate netCDF data format. For example, it used to compute
wind speed and direction which derived from northern and southern wind. The
next step was layer stacking. Both MODIS and weather data were sequentially
stacked to produce the time-series data set and then clipped to cover study area
composite period. The stacked data were evaluated to get temporal pattern from
the time series data (Sakamoto et al. 2005).
Data Analysis
Several points of study area were taken that represented location of paddy
fields. The EVI of these points were analyzed time-series every 16 day during
three years. Weather condition of a pixel weather data where points located was
considered as weather condition of all point of study area because it has coarse
spatial resolution. Weather condition was analyzed time-series daily during 10
years observation. The years observation of EVI data was less than weather data
because the limitation of data storage and for weather study we need long timeseries data.
Identification of Paddy Phenology
We used EVI in our study to detect the phenological stages of paddy. In
order to determine weed control time using spray application, we identified
planting date, heading date, and harvesting date of paddy trough EVI.
Heading time
Heading date used to identify the phenological stages of paddy. According
to time-series data of the spectral reflectance of paddy fields (Domiri et al. 2005),
the maximum EVI appears around the heading date, 60 DAT (Sari et al. 2010;
Semedi 2012). The paddy changes its growth stage from vegetative growth to
reproductive growth on reaching the heading date, and leaves begin to wither and
die. Therefore, we defined the date of the maximum EVI in the time profile as the
estimated heading date and EVI > 0.5 as the threshold.
Planting time
In general, paddy fields are plowed and flooded before planting. The EVI of
paddy fields decreases during this period and then increases again after planting.
Harvesting time
After the heading season, values of EVI begin to decrease as leaves wither
and die and the water in paddy field is dried off before harvesting. EVI then
14
decreases abruptly because of harvesting (Sakamoto et al. 2005). EVI value at the
harvesting stage is around 0.24 – 0.28 (Domiri et al. 2005).
Estimation of Weed Control Time
Normally, weed control in precision agriculture is performed twice, i.e. preplanting and post-emergence (Solahudin 2013). That time estimated after the
planting time known. As mentioned in previous section, paddy phenology from
planting to harvesting represented trough EVI, and the planting time used as
reference to estimate time for weed control. Daily weather condition during preplanting to post-emergence interval was analyzed. Pre-planting estimated a month
before planting month and post-emergence estimated a month after planting
month. Then the interval from pre-harvest to post-emergence considered as weed
control time.
Weed control in paddy cultivation can be performs in various ways. In
Indonesia, generally farmer conducts soil tillage by plowing paddy field before
planting. It purposed to make soil suitable for planting paddy by mixing the
remaining paddy biomass with soil. The paddy field became fertile. It also can kill
the weed. Soil tillage is the first step on weed control in paddy cultivation. Spray
herbicide performed after planting. Time after planting until post-emergence and
it consider as time for spraying herbicide. It is considered the critical period of
weed competition about 40 days after planting.
Development Application to Determine Nozzle Sprayer
This application was developed with objective to determine the proper size
of nozzle sprayer. It expected could minimize drift on weed control. This simple
application could be combined with agents of weed control system (Solahudin
2013) to improve precision of spray application. The rules (IF – THEN) to decide
which nozzle sprayer were acquired from previous researches (Jason 2009;
Solahudin 2013; Tepper 2012) which interpreted into decision tree.
The application was designed and developed using system development life
cycle (SDLC) (O‟Brien 1999). System Development Life Cycle (SDLC) is the
traditional methodology used to develop, maintain, and replace information
system (Hoffer et al. 2007). The phases of the SDLC are: investigation, analysis,
design, implementation and maintenance.
Some rules of decision making based on weather conditions were
(Solahudin 2013):
1) Rules for parameter Wind Velocity (km/h)
- If WV < 2, then do not spray
f 2 ≤ WV ≤ 3, then spraying with a medium droplet and air assist
f 4 ≤ WV ≤ 6, then use a fine droplets
f 7 ≤ WV ≤ 10, then use a medium droplet
f 11 ≤ WV ≤ 14, then use a coarse droplet
f 15 ≤ WV ≤ 20, then spraying with c