IR64 (Indonesia)

 Table 1 summarizes the soil chemical and physical properties of the site.

Based on historical (1953-96) records, the rainy season in Jakenan usually starts in October, peaks in January, and ends in May or June. Average annual rainfall is 1,550 mm, with 1,060 mm falling from November to March. Mean daily solar radiation is low (13 MJ m–2 d–1) in January and high (18 MJ m–2 d–1) in September. Average maximum temperature is 31.7 °C and average minimum temperature is 23.5 °C.

Experimental treatment and design

Field experiments were conducted at two sites, Walik Jerami and Gogonancah, at the Jakenan station in Indonesia from 1995 to 2000. Tables 2 and 3 present the experimental treatments in Walik Jerami and Gogonancah. The treatments were laid out in a split-plot design with four replicates, two water treatments (full irrigation, well-watered; and no irrigation, rainfed) in the main plot and five fertilizer treatments in the 5 × 8-m (40 m2) subplots. In the well-watered treatments, surface irrigation was used to keep the soil at saturated conditions. Plots for the water treatments were lined with polyethylene sheets up to 80-cm depth to minimize subsurface lateral water flow.

Fertilizers were applied in three splits:

 (1) basal: ¼ N, all P, ½ K, all S, and all Zn;

 (2) at maximum tillering: ½ N and ½ K;

 (3) at panicle initiation (PI): ¼ N.

 Micronutrients that included 5 kg Zn ha–1 and 20 kg S ha–1 were applied in all plots. For dry-season crops, basal fertilizer was applied before transplanting. For wet-season crops, basal N was topdressed at 14 days after emergence (DAE) of the dry-seeded crop. The date of fertilizer application in rainfed plots varied within ±7 days around the predetermined dates depending on the occurrence of rainfall to saturate the fields.

Table 2. The experimental design at Gogorancah site at Jakenan station, Indonesia, was a split-plot design with four replicates. The main plot is for water regime and the subplot is for fertilizer level.

Table 3. The experimental design at Walik Jerami site at Jakenan station, Indonesia, was a split-plot design with four replicates. The main plot is for water regime and the subplot is for fertilizer level.

Table 4 shows the experimental treatments for water and tillage management that were used in 1995 and 1996. It was a two-factor experiment (water and tillage) in 1995, whereas three factors (water, tillage, and transplanting date) were involved in the field experiment in 1996. The subplot area was 40 m2.

Table 4. The experimental design at Walik Jerami site at Jakenan station, Indonesia, was a split-plot design with four replicates.

Cultural practices

For all crop seasons, rice cultivar IR64 was planted. For the wet seasons, land preparation started at the onset of the rainy season. All plots were hoed twice manually to 30-cm depth, followed by dry land plowing and harrowing. Five dry rice seeds were dibbled in holes at 15 × 15-cm spacing. After complete emergence, the plants were thinned down to 3 plants per hill. For the dry seasons, hoeing took place right after the harvest of the wet-season crop. Plots were left fallow under submerged or wet conditions for a week, and then harrowed by animal to level the nonpuddled fields and to incorporate basal fertilizer. Three 15- to 21-day-old seedlings were transplanted per hill at 15 × 15-cm spacing. Tables 5 and 6 summarize the planting (seeding or transplanting) and harvesting dates for the six crop seasons.

The farmers’ practices of hand weeding and pesticide spraying were used to minimize pest damage.

Table 5. Seeding, transplanting, and harvest dates for the six seasons of field experiments conducted from 1995 to 2000 at Walik Jerami site at the Jakenan Experiment Station.

Table 6. Seeding, transplanting, and harvest dates for the six seasons of field experiments conducted from 1997 to 1999 at Gogorancah site at the Jakenan Experiment Station.

Plant sampling and nutrient uptake

Dates of emergence, PI, flowering, and physiological maturity were recorded (Tables 5 and 6). At physiological maturity, 22 hills (0.50 m2) were sampled from each subplot for leaf blade, culms and leaf sheaths, panicle dry weights, nutrient (N, P, K) concentrations, and yield components (panicle density, spikelet number per panicle, % of filled spikelets, 1,000-grain weight). Rice yield was determined from a 6-m2 sampling area.

Plant N was determined using Kjeldahl digestion (Varley 1966). The digested solution was analyzed for nitrogen as indo-phenol blue in the Technicon AutoAnalyzer II (Technicon Instruments Corporation 1977).

Water and weather records

Field-water depth was measured daily in 40-cm long, 5-cm diameter PVC tubes installed in each subplot to 25-cm below the soil surface. The bottom 22 cm of the tubes was perforated with 3-mm-diameter holes at 2-cm intervals. Groundwater table depth was measured daily in each main plot of rainfed treatments, using 5-cm-diameter, 150-cm-long PVC tubes, similarly perforated in the bottom 75-cm length, installed to a depth of 100 cm below the soil surface.  In the absence of standing water in the rainfed plots, soil water potentials were measured daily from tensiometers installed at 5-, 10-, and 20-cm depth.

Daily rainfall, solar radiation, maximum and minimum temperature, relative humidity, and wind speed were measured at the Jakenan weather station.

Data analysis

Rice biomass, yield, and yield components were analyzed with standard split-plot analysis of variance techniques.  When the analysis of variance showed significant differences among treatments, Duncan’s multiple range test (DMRT) was used for pair-wise comparison.

In pest-free conditions, grain yields for well-watered NPK treatments were considered as “potential.” Under potential production situations, water and nutrients are nonlimiting so rice growth and yield are determined by weather conditions and crop genetic characteristics only.

Simulation study

The simulations were designed for lowland rice in paddy soil by following the methodology described. Paddy water balance and Penman are employed to calculate the soil water balance and evapotranspiration.  Water and nitrogen production settings are for potential conditions.

Simulations were undertaken for two field experimental sites at two locations, Walik Jerami and Gogorancah, in Jakenan, Indonesia. Tables 6B and 7 present the input files of crop, soil, and the experiment of all simulations. The exact data of crop, soil, and experiment can be easily searched out from the attached database by using the file names listed in Table 6. All simulations were fed by the same weather data from station 43 in Indonesia with the name Indon43.

Table 6B. The input files of simulations for the Walik Jerami site at Jakenan station in Indonesia.

Table 7. The input files of simulations for the Gogorancah site at Jakenan station in Indonesia.

All simulations in Tables 3 and 4 used the same basic crop files (IR64.j96), but the development rates of different phenological stages were parameterized by DRATES and PARAM with the measured data of biomass components. Tables 8 and 9 present the calibrated development rates under potential production conditions at Jakenan station, Indonesia.

Table 8. The development rates of rice plants in various phenological stages for the Walik Jerami site at Jakenan station, Indonesia.

Table 9. The development rates of rice plants in various phenological stages for the Gogorancah site at Jakenan station, Indonesia.

Periodical measurements in the growing season for the depth of ponded water and soil water potential in rainfed fields and aboveground biomass in stem, leaf, and storage organs were used to evaluate the performance of the model in estimating these variables.

Parameter values

Except for the development rates of phenology of rice variety IR64, other parameters were not calibrated in this study. Tables 8 and 9 present the calibrated values of development rates of rice phenology.

Pond water level

Figures 1 and 2 illustrate the temporal changes of modeled and measured depths of ponded water in the rainfed fields at Walik Jerami and Gogorancah at Jakenan station. For experiments at Walik Jerami (Fig.1), the deepest ponded water occurred in 1999 and the shallowest in 2000. The modeled results described the temporal dynamics of water depths, but did not represent the magnitudes of those changes from shallow to deep or from deep to shallow ponded water. The depths of ponded water were underestimated for the early growth season or for fields with moderate water depth. However, the model can correctly estimate the depth of ponded water in the late growth season in the deepest and shallowest ponded fields. 

For the experiments at Gogorancah at Jakenan station, the fields had the deepest ponded water in 1998 and the shallowest ponded water in 1997 (Fig. 2). Again, the temporal sequences of changes in the depth of ponded water were represented by the model. The magnitudes of changes in ponded water were not catched  in 1997 and 1998. The modeled depths explained the measured ones in fields with moderate depths of ponded water in the three seasons.

Regardless of site conditions and management differences, all measured and modeled data pairs were integrated for statistical analysis. Table 10 shows that the depths of ponded water were well simulated by the model. The differences between modeled and measured values were smaller than the variation in measurements.

Table 10. Evaluation results for ORYZA2000 simulations of crop growth and soil water variables over three growing seasons. WAGT, WST, WLVG, and WSO contribute to total aboveground biomass and biomass of stems, green leaves, and storage organs (kg dry matter per hectare). WL is the depth of ponded water (mm). SM is the soil moisture measured (kPa). N is the data pairs of measured and modeled values. MeM and CVm are the mean and coefficient of variation of measurements and MoM is the mean of modeled values.

Soil water content (potential)

ORYZA2000 has predicted the dynamic variations in soil water potential at Walik Jerami and Gogorancah when field water content was lower than field capacity (Figs. 3 and 4). However, the model did not represent the changes in soil water potential from field capacity to saturation and from saturation to field capacity.

Changes in soil water potential from near 0 to lower than –40 kPa (Figs. 3 and 4) were measured two or three times in both field experiments in 1998. These changes possibly resulted from extreme weather conditions, which strongly promoted evapotranspiration. The simulation results did not show these kinds of changes in soil water potential.

Over all the treatments of experiments at the Jakenan experiment station, the model can explain only 28% of the measurements, and the RMSE and RMSEn are larger (Table 10). The low correlation and large RMSE mainly result from the model not capturing the extreme changes in soil moisture (Figs. 3 and 4).

Biomass components

Figures 5 to 12 ( Fig. 5, 6, 7, 8, 9, 10, 11 and 12) illustrate the measured and simulated aboveground total biomass (WAGT), stem biomass (WST), green leaf biomass (WLVG), and storage organ biomass (WSO) in both experiments in different years. The model simulated the temporal variations of all biomass variables under consideration. There was high confidence in the estimations of seasonal yield. Some scatters in graphs implied that the model did not catch all the details of temporal variations within growing seasons.

For goodness-of-fit parameters of biomass components, R2 is significant, and RMSEn is smaller than the CV of measurements (Table 10), indicating a close fit between simulated and measured data. However, the low value of (0.56) for green leaf biomass implies underestimations.

References

Mamaril CP, Wihardjaka A, Abdulrachman S, Suprapto, Fagi AM, Diah WS. 1995. Response of rainfed lowland rice to potassium and sulfur under intensive and diversified cropping systems and low fertility soils. In: Ingram KT, editor. Rainfed lowland rice: agricultural research for high-risk environments. Manila (Philippines): International Rice Research Institute. p 215-225.

Technicon Instruments Corporation. 1977. Technicon AutoAnalyzer II Industrial Method No. 334.74 w/d. In: Individual/simultaneous determination of N and/or P in BD acid digest. Chauncey, New York.

Varley JA. 1966. The determination of N, P, and K ions in plant materials. Analyst 91:119-126.