See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/275575888

Using Crop Simulation Models and GIS for Regional Productivity Analysis

Article  in  Transactions of the ASAE · January 1993

DOI: 10.13031/2013.28328

CITATIONS READS

68 182

5 authors, including:

Harbans Lal Gerrit Hoogenboom

United States Department of Agriculture University of Florida
28 PUBLICATIONS   369 CITATIONS    632 PUBLICATIONS   21,131 CITATIONS   

SEE PROFILE SEE PROFILE

James W Jones
University of Florida
457 PUBLICATIONS   23,368 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Nitrogen Loss and Environmental Assessment Package(NLEAP) View project

Decision Support System for Agrotechnology Transfer (DSSAT) View project

All content following this page was uploaded by Gerrit Hoogenboom on 17 October 2020.

The user has requested enhancement of the downloaded file.

 USING CROP SIMULATION MODELS AND G I S FOR
REGIONAL PRODUCTIVITY ANALYSIS
H. Lai, G. Hoogenboom, J-P. Calixte, J. W. Jones, F. H. Beinroth
MEMBER MEMBER FELLOW
ASAE ASAE ASAE

ABSTRACT. The scope of applicability of site-specific models can be extended to regional planning and productivity
analysis by combining their capabilities with a Geographic Information System (GIS). In this study, regional productivity
analyses for three sites in western Puerto Rico were carried out using the DSSAT-BEANGRO VI.01, the dry bean
(Phaseolus vulgaris L.) model and AEGIS. DSSAT is an integrated decision support system that contains several crop
models with standardized input and output. AEGIS is a regional planning decision support system that uses DSSAT
capabilities within a GIS for regional productivity analysis. The analysis indicated that a considerable soil and weather
variability exists within the three study sites. The optimum management factors such as cultivar selection, planting date
and irrigation strategy would significantly differ from one site to another. For the Mayaguez and Isabela areas, the long
season cultivar 'Porrillo Sintetico* would always be a better choice. It would consistently out perform the short season
cultivar "Cuarentena.*' On the other hand, in the semiarid environment ofMagueyes, 'Porrillo Sintetico* would perform
better than **Cuarentena*' in good rainfall years, and the reverse would happen in bad rainfall years. However, in the
long run, *Porrillo Sintetico' would out yield, *Cuarentena' by 20% even in this area. The simulated **best" planting
dates for rainfed condition matched fairly well with the recommended "best" dates for two study sites. However, for the
third site (Mayaguez), these dates varied considerably for both rainfed and irrigated conditions. The cumulative
probability distribution analysis for these dates showed that the simulated *'best" dates would produce higher yields for
all simulated years. The environmental diversity (soil and weather) of the study sites makes the results of this study
indicative of several other locations in the Caribbean basin and other tropical regions of the world.
Keywords. Crop simulation models. Geographic Information Systems (GIS), Regional analysis and planning. Crop
productivity.

C
rop simulation models predict crop yield, plant Baanante, 1991). Thornton (1991) presented an excellent
growth and development, and nutrient and discussion on how crop models might be implemented at
moisture dynamics based upon site specific farm, field, regional and national levels. Curry et al. (1990)
characteristics such as weather, soil physical and successfully used crop models for studying the possible
chemical parameters, water management, and agronomic effects of climate change on agricultural productivity at the
practices (Whistler, et al., 1986, Penning de Vries et al., national level.
1989). The scope of applicability of these models can be The Decision Support System for Agrotechnology
extended to much broader spatial scales for regional Transfer (DSSAT) integrates several of these models with
planning and policy analysis by combining their standardized input and output (IBSNAT, 1989). Using
capabilities with a Geographic Information System (GIS). DSSAT, various alternatives can be evaluated for a site
Several researchers have demonstrated or discussed the with specific soil and weather parameters to identify the
strengths of this concept for agricultural management most promising crop and management practices. Recently,
decisions and planning at various spatial scales (Curry a regional agricultural decision support system, known as
et al., 1990; Dent and Thornton, 1988; Hoogenboom et al, Agricultural and Environmental (Geographic Information
1990; Hoogenboom and Thornton, 1990; Thornton and System (AEGIS), has been developed (Calixte et al., 1992,
Dent, 1987; Thornton and Dent, 1990; Thornton and Calixte, 1992). AEGIS uses the DSSAT capabilities within
a GIS for regional planning and productivity analysis. In
this combined system, DSSAT provides information on
Article was submitted for publication in January 1992; reviewed and yield and other crop related outputs for different
approved for publication by the Information and Electrical Technologies homogeneous soil and weather combinations, and GIS
Div. of ASAE in December 1992. Presented as ASAE Paper No. 91 -7518. aggregates information from individual units, displays
Approved as Journal Paper No. R-02194 of the Florida Experiment
Station.
maps, and presents results in tabular format for the study
The authors are Harbans Lai, Senior Scientist, ManTech region. AEGIS allows the user to select various
Environmental Technology, Inc., U.S. Environmental Research combinations of crop management practices over space,
Laboratory, Corvallis, OR; Gerrit Hoogenboom, Assistant Professor, evaluate potential crop production, soil loss and other
Dept. of Biological and Agricultural Engineering, The University of factors at a regional scale, and build a land use and
Georgia, Griffin; Jean-Paul Calixte, Graduate Research Assistant,
James W. Jones, Dept. of Agricultural Engineering, University of management plan for the region. A menu-driven interface
Florida, Gainesville; and Fred H. Beinroth, Professor, Dept. of permits the user to perform different processes in AEGIS
Agronomy and Soils, University of Puerto Rico, Mayaguez, Puerto Rico. without having to know the internal structure of the

VOL. 36(1): JANUARY-FEBRUARY 1993 © 1993 American Society of Agricultural Engineers 0001-2351 / 93 / 3601-0175 175
databases and GIS — ARC/INFO (ESRL, 1988) operating 751 mm for the Magueyes Island (fig. 1). Figure 1 also
procedures. The detailed structural and database designs shows monthly averages of total precipitation, daily
and operational functioning of AEGIS are discussed maximum and minimum temperatures, and daily solar
elsewhere (Lai et al., 1990; Lai et al., 1991; and Calixte radiation for all four weather sites. The average daily solar
etal., 1991). radiation varies from 18.57 MJ/m^ at Lajas Substation to
To generate yield databases for AEGIS, we made 22.44 MJ/m2 at Magueyes Island. Historical records for
several thousand simulations with DSSAT for a variety of the weather sites were obtained from Earthinfo Inc. (1989).
management combinations for different soil and weather These records included daily precipitation and maximum
conditions in the study regions. This article analyzes these and minimum temperature. Daily solar radiation values,
simulations to evaluate soil and weather variability, and to required for the DSSAT crop models, were generated using
identify optimum crop management factors such as cultivar the method of Lopez and Soderstorm (1983).
selection, planting dates, and irrigation strategies for the Soil maps (scale 1:20,000) for each study area were
study areas. Regional productivity analyses for the study obtained from soil survey reports of the USDA-Soil
areas were done using AEGIS which helped aggregate Conservation Service (1975). They were digitized into GIS
simulation results from individual soil mapping units to (pc ARC/INFO). Sixty-six soil mapping units suitable for
produce tabular reports and regional maps. agriculture occur in the three areas. Tliey belong to alfisols,
entisols, inceptisols, mollisols, oxisols, ultisols, and
vertisols; and thus exemplify seven of the eleven orders
METHODS recognized in the U. S. system of soil classification. Soil
STUDY AREAS CHARACTERISTICS AND DATA mapping units are characterized by their series name,
DEVELOPMENT slopes, and their erodibility status. Soil and climatic
Three areas of about 3800 ha (9400 acres) each located diversity within the three areas make them good candidates
near Mayaguez, Isabela, and Magueyes were selected in for the study. Their performance would be indicative of
western Puerto Rico. The climate in these areas range from several other locations in the Caribbean basin and other
humid to subhumid to semiarid tropical. The Mayaguez tropical regions of the world.
region with mean annual precipitation of 1950 mm is To facilitate analysis and presentation of results at
characterized as humid, Isabela as subhumid with mean regional scales, the agriculturally suitable mapping units
annual precipitation of 1549 mm, and Magueyes as within the study areas were categorized into 11 groups.
semiarid tropics. For the Magueyes area we selected two These groups, developed by an experienced soil scientist of
weather sites, namely, Lajas Substation and Magueyes the region, contained mapping units of similar properties
Island, because a mountain belt divides the region into two and productivity level. Table 1 presents distribution of
distinct climatic zones. The average annual precipitation these groups within the four weather sites and table 2
for these sites are 1137 mm for Lajas Substation and shows mapping units combined in each group. Spatial

10 11 12

300
Magueyes Island
250 ^

or-"^ -e^DaHy Max.Temp. — D a i y MIn.Temp. ^ E

200 ^-^
^DdfySolwRad. D Monthly Prwdp.
c
.2
150 - g

100 g
Q.
50

1 2 3 4 5 6 7 8 9 10 11 12
Month
Figure 1-Monthly averages of total precipitation, and daily maximum temperature, minimum temperature, and simulated solar radiation for
the four weather sites in the study areas of Puerto Rico.

176 TRANSACTIONS OF THE A S A E

 Table 1. Distribution of soil groups under Table 2. Generic soil names and associated mapping
four weather regions in Puerto Rico units in different soil groups
Weather Locations Group
No. Group Name Series Name Mapping Units
Lajas Magueyes
Mayaguez Isabela Substation Island 1 Deep, clayey soils Aceitunas AaC2
of undulating Consumo CoE, CoF2
Soil Area Area Area Area humid uplands Humatas HmD2, HmE2
Group* (ha) % (ha) % % (ha)^ % Lares LaB2,LaD2
1 121.5 23.72 6.6 0.22 0.0 0.00 0.0 0.00 Montegrande Mvc
2 43.5 1.42 0.0 0.00 0.0 0.00 0.0 0.00 Morado MwE2
3 304.7 9.93 0.0 0.00 0.0 0.00 0.0 0.00 Mucara MxD2,MxE2
4 0.8 0.03 0.0 0.00 0.0 0.00 0.0 0.00 Nipe NcD2
5 0.0 0.00 1637.3 54.05 0.0 0.00 0.0 0.00 Rosario RsD2, RsE2
6 0.0 0.00 317.0 10.47 0.0 0.00 0.0 0.00
7 0.0 0.00 325.9 10.76 0.0 0.00 0.0 0.00 2 Shallow, clayey Mucara MxC, MxF2
8 0.0 0.00 0.0 0.00 299.4 20.95 549.3 22.70 soils of steep Malaya McF2
9 0.0 0.00 0.0 0.00 164.9 11.54 155.5 6.43 humid uplands Quebrada QuE2
10 0.0 0.00 531.8 17.56 91.1 6.38 1087.3 44..93 Morado MwF2
11 0.0 0.00 0.0 0.00 849.0 59.41 72.4 2.99
NS/SEA 1990.7 64.91 210.9 6.96 24.6 1.74 555.8 22.97 3 Deep, clayey and Bajura Ba
loamy soils of humid Coloso Cn
Total 3067.2 100.00 3029.5 100.02 1429.0 100.00 2420.3 100.00 floodplains Toa ToA
* Soil Groups
1 Deep, clayey soils of undulating humid uplands. Shallow, gravelly Reilly Re
2 Shallow, clayey soils of steep humid uplands. soils of humid uplands
3 Deep, clayey and loamy soils of humid floodplains.
4 Shallow, gravelly soils of humid uplands. 5 Deep, clayey soils of Goto CuB2, CvB
5 Deep, clayey soils of subhumid plains. subhumid plains Cotito CtB2
6 Deep, sandy soil of subhumid plains. Matanzas BsB, MtB
7 Shallow, clayey soils of subhumid upland.
8 Shallow, clayey soils of steep semiarid uplands. 6 Deep, sandy soil of Bejucos BeB
9 Deep, gravelly soils of steep semiarid uplands. subhumid plains Guerrero GuB
10 Shallow, gravelly soils of semiarid uplands. Jobos JoB
11 Deep, clayey soils of semiarid flatplains. Rio Lajas RiB
NS/SEA Soils not suitable for Agriculture/Area under sea.
7 Shallow, clayey soils Colinas CiB
of subhumid uplands Solter SrD
pattern of soil groups in the Magueyes region is shown in Santa Clara SeB
figure 2. Tanamara TcB2,TcC2,TcD2

MODEL SELECTION AND SIMULATIONS 8 Shallow, clayey soils Descalabrado DeC, DeD, DeF
The DSSAT dry bean (Phaseolus vulgaris L.) model of step semiarid Jacana JaB, JaC,
(BEANGRO) was used to simulate yield and irrigation uplands JaC2,JaD2
requirements for a variety of management, soil, and
9 Deep, gravelly soils Amelia/Maguayo AmB, AmC2
weather combinations. Dry bean is a staple food crop for
of semiarid uplands Jacana JcB, JcC
the region. The current large import of dry bean into Puerto
Rico, and the possibility of replacing sugarcane in some 10 Shallow, gravelly soils San German SaD, SaE, SgD, SgF
areas because of economical reasons, make this crop a of semiarid uplands Guayama GuD,GuF
good candidate for the study. Pozo Blanco PoA, PzB, PzC, PzD
BEANGRO Vl.Ol, though the youngest model in the
DSSAT-GRO series, has been calibrated and tested for 11 Deep, clayey soils of Aguirre Ag
several locations with promising results. It has been semiarid flatplains Cartagena Ca
calibrated and tested for Gainesville, Florida; Guatemala; Fratemidad FrA, FrB, FvA, FvB, Ft
and Cali, Colombia, with a variety of management
conditions including Torrillo Sintetico' a black-seeded
long-season cultivar commonly grown in Central and to crop growth and yield. Even with these limitations, the
South America (Hoogenboom et al., 1990a; Hoogenboom model results can be very effective in evaluating spatial
et al., 1990b; Thornton and Hoogenboom, 1990). and temporal variability in crop production due to soil
BEANGRO was also able to predict yields for a white- types, weather, and management factors.
seeded short season cultivar 'Cuarentena' within the range The DSSAT soil datafiles for the agricultural suitable
of 10 to 15% in two experiments conducted in Isabela, mapping units in the study areas were created from the
Puerto Rico (table 3). For the present study, we used both USDA-SCS soil characterization database using DSSAT
of these cultivars 'Cuarentena' and Torrillo Sintetico.' (IBSNAT, 1989). The parameters for the soils not available
BEANGRO like other DSSAT-GRO series models assume in the USDA-SCS databases were estimated using soil
that nutrients are not constrained, and pests (insects, taxonomic analogies and expert knowledge (Beinroth,
diseases, and weeds) are controlled and pose no limitations 1990).

VOL. 36(1): JANUARY-FEBRUARY 1993 177

 3
Porrillo Sintetico The 20-year simulations were analyzed to generate
average, 20 and 80-percentile values of output factors
- (yield, irrigation requirement, etc.) for each management
^^„^
2.5 ppi i .
iil
i i l iiii i iil ] fifca i T T T T

combination (planting date, cultivar, and irrigation type)

1! l i l i y i i
(0 2
- Mmm i;| m fmm f under different soil and weather conditions. The 20 and
€ iiii 111 I i i i i li pilii 1f i i
C i | pq:*?*! |: ipH* 80-percentile values are the 4th and 16th values in a set of
g l . 5 - liiiii III III 111 ii|ii| 1i l 20 observations arranged in ascending order. In statistical
•o il i i i i 1 terms these percentiles indicate the probability of getting
•^ 1
III 1 iiii 1i l l
> ^ 1 III 1
mil less than or equal to a certain value of the parameter being
- analyzed. For example, the 20-percentile value indicates
0.5 III III 1
that 80% of the time one would expect to get yield equal or
0
less than the quoted yield value. Further analysis of these
simulations was carried for evaluating spatial and temporal
Cuarentena variability in crop production due to soil types, weather,
3
H i Average Yield I 20 To 80 Percentile Variation
and management factors that could be used for generating
2.5 management and planning recommendations for the
regions.
^i_^
2 - pii l i i i i l i i l f i pEi] ][ i i l i jX T
m
T T

CO i i i i i pf
fi.5
o
iiii
li
11 p i l l 1i i l
RESULTS AND DISCUSSION
Regional productivity analysis involves evaluating
^^ 1
-o iiii II [ i i i i 1 l i i
Q> biiiil m i i i i ii i i spatial soil and weather variability, identifying optimum
> iiii iiii li iiii i i l l crop management practices, and predicting productivity of
0.5
III the region under different climatic and management
Iiii ill III scenarios. These analyses can help regional planners and
CoF2 HmE2 LaD2 MwE2 MxE2 RsD2 policy makers in delineating acreage and distribution of
CoE HmD2 LaB2 MvC MxD2 NcD2 RsE2
areas with high productivity (soil and climate) and
Soil Mapping Unit developing management recommendations for different
Figure 2-Distributioii of soil groups in Magueyes region. crops.

Simulations for the four weather sites over 20-historical

weather years (1961-1980) were made using the Strategy Porrillo Sintetico
Evaluation program of DSSAT for all agriculturally
suitable soil mapping units. The management combinations r4-| r ^
simulated were 12 planting dates (15th day of each month),
4 ™ . rl^
rh r^ Hh
2 cultivars (short season and long season), and 2 irrigation CO 2

types (rainfed and irrigated). For the irrigated treatment, c

water was applied to fill the profile whenever the available S,1.5
moisture dropped to 50% in the upper 30 cm. Row spacing J-
and plant density were kept uniform at 0.60 m and -
13 plants/m^ for all simulations. A report that consisted of
input factors including planting date, irrigation strategy,
and cultivar; and the simulated output factors including
seed yield, biomass, evapotranspiration, number and
amount of irrigations, nitrogen leached from the profile, Cuarentena
and days to different levels of canopy cover were saved for
each simulation.

Table 3. Climatic parameters, measured, and

BEANGRO predicted yields for two
experiments at Isabela, Puerto Rico
Year Year
Parameters 1989 1990
Tmax C C) 29.9 30.4
T m i n r C) 20.1 21.1
Rain (mm) 242 450
Solar Rad. (MJ) 1263 1553 CoF2 HmE2 LaD2 MwE2 MxE2 RsD2
CoE HmD2 LaB2 MvC MxD2 NcD2 RsE2
Cultivar Cuarentena Cuarentena
Soil type Deep clay Deep clay Soil Mapping Unit
Measured yield (kg/ha) 988 1289 Figure 3-Simulated yield variation for dry bean planted on
BEANGRO predicted 15 September under rainfed conditions for different mapping units
yield 1123 1325 within Soil Group 1 (deep, clayey soils of undulating humid uplands)
for Mayaguez weather in Puerto Rico.

178 TRANSACTIONS OF THE A S A E

SOIL AND WEATHER VARIABILITY variability. On the other hand, the semiarid climate of
The soil variability is mainly restricted to spatial level, Magueyes Island produced low yields with high variability
while weather can vary over space and time in a region. for all soil groups (8, 9, 10, and 11). The shallow gravelly
For evaluating the soil variability, we analyzed soil soils of semiarid uplands (Soil Group 10) performed the
mapping units within different groups and soil groups poorest under both semiarid and subhumid climates of
within different climatic zones. The spatial weather Magueyes and Isabela regions, respectively. This was
variability was analyzed based upon the yield response of a largely because of low water holding capacities of these
soil group under two or more climatic zones, and the soils.
temporal weather variability was based upon year-to-year Yield differences between Lajas Substation and
variations at a particular site. For the tropical climate of Magueyes Island climates for Soil Groups 8, 9, 10, and 11
Puerto Rico, all these variabilities can significantly affect (fig. 4), are examples of spatial weather variability. All
crop yields and the optimum crop management factors. these soils produced higher yields under the Lajas
Figure 3 presents the mean yield values along with 20 to Substation climate as compared to the Magueyes Island
80-percentile variations for the both simulated cultivars climate, thus proving the Lajas Substation climate to be a
planted on 15 September under rainfed condition for more favorable environment for bean production. The
different mapping units within Soil Group 1 (deep, clayey effect of temporal variability for the Mayaguez and
soils of undulating humid uplands). Most mapping units in Magueyes weather sites is presented in figure 5. It presents
this group produced fairly stable yields with high averages yields for the two cultivars ('Cuarentena' and 'Porrillo
(approximately 2.0 ton/ha for Cuarentena, 2.5 ton/ha for Sintetico') planted on 15 September under rainfed
Torrillo Sintetico') and low variability. This validates the conditions for 20 consecutive years (1961-1980). Yields
soundness of the criteria employed for combining similar were consistently higher with low year-to-year variation
mapping units into a single soil group. under the Mayaguez climate for both cultivars. This
Yield responses of soil groups found in different establishes stability of the Mayaguez climate for bean
climatic zones of the study areas are presented in figure 4. production.
This figure presents average yields along with 20 to
80-percentile values for 'Cuarentena' dry bean planted on CROP MANAGEMENT FACTORS
15 September under rainfed conditions. The average yield Proper selection of cultivar(s) and other management
varied from 0.23 ton/ha for Soil Group 10 (shallow factors can considerably enhance regional productivity
gravelly soils of semiarid uplands) in the Magueyes Island without increasing risks associated with them. The two
climate to 2.13 ton/ha for Soil Group 3 (deep, clayey, and cultivars simulated in this study responded differently to
loamy soils of humid floodplains) under the Mayaguez different soil and weather conditions of the study areas.
climate. All soil groups ( 1 , 2, 3, and 4) under the The growing season for 'Cuarentena' varied between
Mayaguez climate produced high yields with low 66 and 68 days, and that of 'Porrillo Sintetico' between

Mayaguez Isabela
3 3
i Average Yield X 20 To 80 Percentile Variation
Soil Group
2.5 2.5
1. Deep, clayey soils of
undulating humid uplands
1? 2 2. Shallow, Clayey soils of 2
steep humid uplands rm
§1.5 by] 3. Deep, clayey and loamy 1.5
r 1 soils of humid floodplains
Jl L \1 4. Shallow, gravelly soils
of humid uplands
1
^ii
V "^1

0.5 0.5

0 0
1 2 3 8 9 10 11 1 2 3 4 8 9 10 11
Lajas Substation Magueyes Island
3 3
Soil Group Soil Group
2.5 5. Deep, clayey soils of 2.5 9. Deep, gravelly soils
subhumid plains of semiarid uplands
r 6. Deep, sandy soil of 2 10. Shallow, gravelly soils
of semiarid uplands
subhumid plains 11. Deep, clayey soils of
T
7. Shallow, clayey soils 1.5 semiarid flatplains
of subhumid uplands
T
5= 1 [• 8. Shallow, clayey soils 1
of steep semiarid uplands
0.5 0.5

0
5 6 7 8 9 10 11
0
1 5 6 7
MJjjb.
8 9 10 11
Soil Group Soil Group
Figure 4~Simulated dry bean yield response of soil groups found under different climatic zones of the study areas in Puerto Rico when planted
with Tuarentena' on September 15 under rainfed condition.

VOL. 36(1): JANUARY-FEBRUARY 1993 179

 74 and 76 days for all simulations. Performance of the two
3 cultivars on the most predominant soil mapping unit for
Mayaguez (CoE), Sept. 15, Rainfed each of the four weather sites is presented in figure 6. This
2.5 figure shows cumulative probability functions for the
simulated yields with the two cultivars planted on
«" 2
15 September (the recommended "best" planting date)
under rainfed conditions for 20 weather years. Under the
humid climate of Mayaguez and the subhumid climate of
> 1 Isabela, the long season cultivar, Torrillo Sintetico'
0.5
performed consistently better than the short season cultivar
'Cuarentena' and should be the obvious choice for these
environments. However, it is more difficult to select a
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 cultivar for the semiarid environments of Magueyes Island
Year Of Simulation
and Lajas Substation. For both of these locations, the
performance of Torrillo Sintetico' almost equaled that of
3 'Cuarentena' at lower yield levels representing low rainfall
Magueyes island (JcB), Sept. 15, Rainfed
years. At higher yield levels, representing good rainfall
2.5 years, the cultivar of 'Porriilo Sintetico' yielded
considerably higher than 'Cuarentena'. Thus, the long
'(0 2 I I Cuarentena
season cultivar, 'Porriilo Sintetico', also seemed to be a
"c • 1 Porriilo Sintetico
better choice for the semiarid environment of the

u I1
Magueyes region. To test this hypothesis, we used AEGIS
to carry out a spatial analysis and aggregation of yields

MM a
from different mapping units for the two climatic zones,
different soil groups, and the entire region.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
n. Yield distribution maps for the two cultivars planted on
Year Of Simulation
15 September for the Magueyes region under rainfed
conditions are presented in figure 7. The northern part of
Figure 5-EfFect of temporal weather variability on simulated dry the region is simulated using the Lajas Substation weather
bean yield for two sites in the study areas of Puerto Rico.
and the southern part using the Magueyes Island weather

Mayaguez (CoE2), Sept. 15, Rainfed Isabela (AaC2), Sept. 15, Rainfed
f 9
0.95 I / 0.95
^* -H^ Cuarentena
= 0.8 { f 0.8 ^Porriilo Sintetico
CO \ \
-90.65 iI }I 0.65
O
u. f f
I I
Q-0.5
0) iT itt 0.5
> T
••50.35
CD
3
i 0.35

£ 0.2
D
J
/i
J
il 0.2
Oo.05 .^=^^^— 0.05
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5

Lajas Substation (JcB), Sept. 15, Rainfed Magueyes Island (JcB), Sept. 15, Rainfed
0.95

1 1.5 2
Yield (ton/ha) Yield (ton/ha)
Figure 6-Simulated dry bean yield response of long season, *Porrillo Sintetico', and short season, 'Cuarentena', cultivars from the most
predominant soil mapping unit planted on 15 September under rainfed condition, under different weather sites in the study areas of Puerto
Rico.

180 TRANSACTIONS OF THE A S A E

 Table 5. Productivity of different soil groups and overall yield for
Average Yield f o r B e a n s (SS) roinfed. Lajas Valley region for Torrillo Sintetico' and 'Cuarentena'
Mogueyes Sep 15 - Nov 24 Yield per ho
planted on 15 September under rainfed conditions
I I not cropped
Lajas Substation Magueyes Island
Q <= 0.40 metric ton

g <= 0.80 metric tons

Soil Area Production (tons) Area Production (tons)
^ <= 1.20 metric tons

^ <= 1.60 metric tons

Group* (ha) (ha)
^ <= 2.00 metric tons
Cuarentena Porrillo Cuarentena Porrillo
8 299.4 355.0 416.6 549.3 230.1 264.6
•I > 2.00 metric tons
9 164.9 282.5 342.7 155.3 158.7 184.4
10 91.1 49.4 56.6 1087.3 152.3 198.4
11 849.0 1224.3 1575.5 72.4 56.4 71.6

Total 1404.4 1911.2 2391.3 1864.3 597.5 714.0

Avg. yield
(tons/ha) 1.36 1.70 0.32 0.38
A v e r a g e Y i e l d f o r Beans (LS) roimed.
Mogueyes Sep 15 - Dec 4 yield per ho * Soil group
r n not cropped
8 Shallow, clayey soils of steep semiarid uplands.
^ <= 0.40 metric ton

3 <= 0.80 metric tons

9 Deep, gravelly soils of semiarid uplands.
J <= 1.20 metric tons 10 Shallow, gravelly soils of semiarid uplands.
^ <= 1.60 metric tons 11 Deep, clayey soils of semiarid flatplains.
I <= 2.00 metric tons

I > 2.00 metric tons

'Cuarentena'. This establishes the superiority of the

Torrillo Sintetico' cultivar even for the Magueyes region.
However, the overall low yields under Magueyes Island
climate can mainly be attributed to its low annual
precipitation (751 mm) as compared to 1137 mm for Lajas
Figure 7-Yield maps for the Magueyes region in Puerto Rico based Substation.
upon BEANGRO simulations for two dry bean cultivars Torrillo Planting date is another important management factor
Sintetico' and ^Cuarentena', planted on 15 September under rainfed that significantly influences regional productivity. Figure 8
conditions. presents simulated mean yields and irrigation applied to
Torrillo Sintetico' when planted on different dates under
records. Areas under different yield ranges, and the rainfed and irrigated conditions for the most predominant
production from different soil groups, also generated by soil group of each climatic zone. Under rainfed conditions,
AEGIS, are presented in tables 4 and 5. Total productions the highest yielding planting dates varied considerably
for different soil groups and from the entire region were from one region to another. For the semiarid climates of
estimated by calculating production (yield*area) from all Lajas Substation and Magueyes Island, the plantings in
individual polygons with homogeneous characteristics (soil August, September, and October produced the highest
and weather) and aggregating them for the entire area. yields. The lowest yield for these sites occurred with
Areas under higher yield ranges increased for both January and May plantings. The low yields were primarily
climatic zones when planted with Torrillo Sintetico' attributed to high temperatures coupled with low
(table 4). The overall productivity (total production/total precipitation during the cropping period. In the subhumid
area) increased from 0.32 to 0.38 tons/ha in the Magueyes and humid climates of Isabela and Mayaguez, the May
Island climatic zone, and from 1.36 to 1.70 tons/ha in the planting produced reasonably good yields, although the
Lajas Substation climatic zone (table 5). It represented an highest yielding planting dates for these sites were 15 April
increase of approximately 20% in production for both areas and 15 June, respectively.
when planted with 'Porrillo Sintetico' instead of Irrigation dampened the planting date effect. It
considerably reduced the yield differences among the four
weather sites. However, the irrigation requirement varied
Table 4. Distribution of areas under different yield ranges for
Lajas Valley region under rainfed conditions for short- significantly under the four environments. It decreased
^Cuarentena' and long Torrillo Sintentico' season with an increase in rainfall - from semiarid to subhumid to
dry bean planted on 15 September humid tropics. The semiarid climate of the Magueyes
Yield Range Area (ha) Island required considerably higher irrigation (130 to
(tons/ha) 245 mm) than the humid climate of Mayaguez (2 to
Lajas Substation Magueyes> Island
50 mm). Early planting dates (January to March)
Min Max Cuarentena Porrillo Cuarentena Porrillo responded more favorably to irrigation than the later ones
0.0 0.4 51.75 0.00 1342.53 1339.51 in all regions. However, the mid-year planting dates (June,
0.4 0.8 107.75 135.59 366.44 297.10 July, and August) did not produce any yield advantage
0.8 1.2 44.33 42.64 155.51 169.48 under the humid climate of Mayaguez. The yield variation
1.2 1.6 1035.80 41.27 0.00 58.39 under irrigation for different planting dates, especially
1.6 2.0 164.95 1020.08 0.00 0.00 under Mayaguez and Isabela climates, followed a
>2.0 0.00 164.95 0.00 0.00
sinusoidal curve similar to the variation in mean monthly
1404.58
temperature (fig. 1) with a phase shift of 90°. This
Total 1404.53 1864.48 1864.48
indicates that ambient temperature in addition to soil

VOL. 36(1): JANUARY-FEBRUARY 1993 181

 Mayaguez, Soil Group 1, Porrillo Sintetico Isabeia, Soil Group 5, Porrillo Sintetico

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN
ririnnnrannnn
FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Planting Month Planting Month

Lajas Substation, Soil Group 11, Porrillo Sintetico Magueyes Island, Soil Group 10, Porrillo Sintetico
250 I =• • ] , 4

*J
t*4 (0

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Planting Month Planting Month
Figure ^Effect of planting dates on mean yields for a long-season 'Porrillo Sintetico' dry bean cultivar under rainfed and irrigated conditions,
and irrigation applied for the most predominant soil group under each weather site in the study areas of Puerto Rico.

moisture becomes a limiting factor for irrigated crop planting date, and irrigation strategy; and estimate
production under the tropical climate of Puerto Rico. productivity of the study areas under different management
Under rainfed conditions, the simulated "best" planting scenarios.
dates matched with the currently practiced and A considerable soil and weather variability exists within
recommended dates by local extension specialists (table 6 the three study sites. All soils within the Mayaguez region
and fig. 8) for two of the three study areas. For the third produced higher yields with low variability because of
site, 'Mayaguez', these dates varied considerably from the productive soils coupled with its favorable climate of
recommended dates both for rainfed and irrigated humid tropics. On the other hand, the semiarid climate of
conditions. For the rainfed condition, the recommended Magueyes region produced low yields with high variability
dates were 15 September or 15 April, and the simulated for all soil groups. The analysis helped identify the
"best" date was June 15; and for the irrigated condition the optimum management packages for the three sites. For the
simulated "best" planting was 15 February instead of Mayaguez and Isabeia areas, the long season cultivar
15 November. A cumulative probability distribution 'Porrillo Sintetico' would always be a better choice. It
analysis for these planting dates (fig. 9) showed that the would consistently out perform 'Cuarentena' - the short
simulated "best" planting dates would consistently produce season cultivar. In the semiarid environment of Magueyes,
higher yields than the recommended dates for all years 'Porrillo Sintetico' would perform better than 'Cuarentena'
both for rainfed and irrigated conditions. A long-term field in good rainfall years, and the reverse would happen in bad
experiment to test these results is recommended.
Table 6. Recommended and simulated '^best" planting
SUMMARY AND CONCLUSIONS dates in the three study areas in Puerto Rico
Regional productivity analyses for three sites located
Irrigation "Best" Planting Date (s)
near Isabeia, Mayaguez, and Magueyes in western Puerto
Rico were carried out using the DSSAT-BEANGRO Study Area Management Recommended* Simulated
VI.01, the dry bean (Phaseolus vulgaris L.) model, and Mayaguez Rainfed 15 Sept./15 April 15 June
AEGIS. DSSAT is an integrated decision support system Isabeia Rainfed 15 April 15 April
that contains several crop models with standardized input Magueyes Rainfed 15Sept./15 0ct. 15 Sept./15 Oct.
and output. AEGIS is a regional planning decision support
system that uses DSSAT capabilities within a GIS for Mayaguez Irrigated 15 November 15 February
regional productivity analysis. The analysis of results of Isabeia Irrigated 15 November 15 April
several thousand simulations with BEANGRO helped Magueyes Irrigated 15 November 15 March/15 July
evaluate soil and weather variability; identify optimum Courtesy of Dr. Jim Beaver, Bean Breeder, University of Puerto
crop management practices such as cultivar selection, Rico, Mayaguez, Puerto Rico.

182 TRANSACTIONS OF THE A S A E

 Mayaguez (MxE2), Porrillo Sintetico REFERENCES
Beinroth, F. H. 1990. Estimating missing soil data: How to cope
Irrigated with an incomplete database. Agrotechnology Transfer
Feb. 15 (Simulated Best)
o 0.85 11:10-11.
c 0.75 Nov. 15 (Recommended)
3 Calixte, J. P. 1992. Design and development of a regional
c agricultural decision support system. M.S. thesis, Agricultural
0) 0.65

^< 0.55 Engineering Dept., University of Florida, Gainesville.

-n 0.45
CD Calixte, J. P., H. Lai, J. W. Jones, F. H. Beinroth and L. R. Perez-
o Alegria. 1991. A model based decision support system for
or 0.35 regional agricultural planning. Proceedings of Conference on
^D- New Directions in Sustainable Agriculture, 14-19 July,
Caribbean Food Crops Society, Roseau, Dominica.
Calixte, J. P., F. H. Beinroth, J. W. Jones and H. Lai. 1992.
Linking DSSAT to a geographic information system.
Agrotechnology Transfer (15): 1-7.
Curry, R. B., R. M. Peart, J. W. Jones, K. J. Boote and L. H. Auen,

(I
Jr. 1990. Simulation as a tool for analyzing crop response to
0.95 Rainfed climate change. Transactions oftheASAE 33(3):981-990.
Q 0.85 -^ Jun. 15 (Simulated Best)
Dent, J. B. and P. K Thornton. 1988. The role of biological
C *- Sep. 15 (Recommended)
3 0.75 simulation models in farming systems research. Agricultural
c ' Administration and Extension 29:111 -122.
0) 0.65
Earthlnfo, Inc. 1989. Climatedata, Vol. 3.2, Summary of the day -
CD 0.55 -
i i Eastern, NOAA-NCA, Boulder, CO.

to

=
0.35

0.25
-
1 /
^^^^--^ /
ESRI. 1988. pcARC/INFO Starter Kit. Environmental Systems
Research Institute, Inc. Redlands, CA.
Hoogenboom, G. and P. K. Thornton. 1990. A GIS for
agrotechnology transfer applications in Guatemala. In
0.15
^^^^ } Application of GIS, Simulation Models, and Know ledge-based
0.05 m ^
/ Systems for Landuse Management, 61-70. Agricultural
0.5 1 1.5 2 2.5 3 3.5 Engineering Dept., Virginia Tech., Blacksburg, VA.
Yield (ton/ha) Hoogenboom, G., C. E. Heer and P. K. Thomton. 1990. Farm
Figure 9-Cuinulative probability distribution analysis for simulated modeling in Guatemala. Agrotechnology Transfer 12:6-12.
"best" and recommended "best" planting dates for dry bean 'Porrillo Hoogenboom, G., J. W. White and J. W. Jones. 1990a. A computer
Sintetico' in the Mayaguez region of Puerto Rico. model for the simulation of bean growth and development. In
Advances in Bean (Phaseolus vulgaris L.) Research and
rainfall years. However, in the long run 'Porrillo Sintetico' Production, Centro Intemacianal de Agricultura Tropical, Cali,
would out yield 'Cuarentena' by 20% even under the Columbia.
climate of Magueyes Island with annual precipitation of Hoogenboom, G., J. W. Jones and K. J. Boote. 1990b. Modeling
751 mm. growth, development, and yield of legumes: Current status of
SOYGRO, PNUTGRO, and BEANGRO models. ASAE Paper
Under rainfed conditions, the optimum planting dates
No. 90-7060. St. Joseph, MI: ASAE.
and yield levels varied considerably from one region to IBSNAT. 1989. Decision Support System for Agrotechnology
another. The simulated "best" planting dates matched fairly Transfer (DSSAT)-User's Guide. IBSNAT Project, Dept.
well with the recommended "best" dates for two study Agronomy and Soil Sci., University of Hawaii, Honolulu.
sites. For the third site, 'Mayaguez', these dates varied Lai, H., J. W. Jones, F. H. Beinroth and R. M. Peart. 1990.
considerably for both rainfed and irrigated conditions. The Regional agricultural planning information and decision
cumulative probability distribution analysis for these dates support system. In Application of GIS, Simulation Models and
showed that the simulated "best" dates would produce Knowledge-Based Systems for Landuse Management, 51 -60.
higher y i e l d s for all simulated years. Irrigation Agricultural Engineering Dept., Virginia Tech., VA.
considerably dampened the planting date effects. It Lai, H., J. P. Calixte, J. W. Jones, F. H. Beinroth and L P. Alegria.
indicates that with irrigation, dry bean can be grown 1991. Overview of agricultural and environmental geographic
information system (AEGIS). Proceedings of the National Soil
around the year at all three sites. However, the irrigation
Survey Cooperative Conference, 33-38. Seattle, WA.
requirement decreased with the increase of rainfall - from
Lopez, A. M. and K. G. Soderstorm. 1983. Insolation in Puerto
semiarid to subhumid to humid tropics. Early planting Rico. In Solar Engineering-1983, ed. L. M. Murphy, 70-75.
dates (January to March) responded more favorably to New York: American Society of Mechanical Engineers.
irrigation than the later ones in all regions. A long-term Penning de Vries, F. W. T., D. M. Jansen, H. F. M. ten Berge and
field experiment to test these results is recommended. A. Bakema. 1989. Simulation of Ecophysiological Processes of
The study successfully demonstrated that the scope of Growth in Several Annual Crops. The Netherlands: Pudoc,
applicability of site-specific models can be extended to Wageningen.
regional planning and productivity analysis by combining Thomton, P. 1991. Application of crop simulation models in
their capabilities with GIS. These combined tools can be agricultural research and development in the tropics and
subtropics. International Fertilizer Development Center,
used for characterizing the region, locating high yielding
Muscle Shoals, AL (Paper series: P-15).
and problematic areas, and estimating their productivity
Thomton, P. K. and J. B. Dent. 1987. IBSNAT crop models in a
under different management scenarios. socio-economic whole farm framework. Agrotechnology
Transfer 6'A'l.

VOL. 36(1): JANUARY-FEBRUARY 1993 183

 . 1990. Using models to explore farming alternatives. USDA-Soil Conservation Service. 1975. Soil Survey ofMayaguez
Agrotechnology Transfer 11:1-9. Area of Western Puerto Rico. Washington, DC: GPO.
Thornton, P. K. and C. A. Baanante. 1991. Use of simulation Whisder, F. D., B. Acock, D. N. Baker, R. E. Frye, H. F. Hodges,
models for evaluating sustainable production systems. J. R. Lambert, H. E. Lemmon, J. M. McKinnion and V. R.
Presented at the Symposium on Sustainability of pasture crop Reddy. 1986. Crop simulation models in agronomic systems.
rotations in the Southern Cone of South America. Montevideo, Advances in Agronomy 40:141 -208.
Uruguay, 27-29 May.

184 TRANSACTIONS OF THE ASAE

View publication stats