Table of Contents

1. Introduction
2. Simulation platforms
   1. ORYZA
      • Prepare for the inputs
      • Scenario
      • Run the simulation
   2. GAMA
      • Format processing
      • daily_result model
      • seasonal_result model
      • eco model

Introduction

This repo combines crop simulation modeling with agent-based modeling to understand both the rice performance and socio-economic adoption patterns of different water management strategies in rice farming.

In this experiment, I want to compare 2 water management practices (Continuous Flooding & Alternate Wetting and Drying) to see how it affects on the rice yield, water usage, emissions. And then a farmer decision-making model will simulate the decision of farmer whether to choose CF or AWD based on the economic aspects. 2 Irrigation practices for comparison

1. Continuous Flooding (CF): The traditional method where rice fields (paddies) are kept continuously flooded with 5-10 cm of standing water throughout most of the growing season

   • Advantages:
     • Familiar to farmers (traditional practice)
     • Suppresses weeds effectively
     • Generally stable yields
     • Lower risk of crop stress
   • Disadvantages:
     • Very high water consumption (can use 1,500-2,000 mm of water per season)
     • Higher methane emissions
     • Not sustainable in water-scarce regions

2. Alternate Wetting and Drying (AWD): A water-saving technique where irrigation water is applied to flood the field, then allowed to dry down to a certain level (usually 15 cm below soil surface) before re-flooding

   • Advantages:
     • Saves 15-30% water compared to CF
     • Reduces methane emissions
     • Can maintain yields if managed properly
   • Disadvantages:
     • Requires more careful monitoring
     • Risk of yield reduction if soil dries too much
     • Farmers may be hesitant to adopt (unfamiliar practice)
     • Requires proper drainage infrastructure

I compared double-rice season (Winter-Spring and Summer-Autumn) with 2 water management practices under 3 scenarios from 2015 to 2020 in Kien Giang (old)

• Baseline (in normal condition)
• Temperature rises 2℃
• AWD Threshold Changing (let the field drain more)

Simulation platform

ORYZA

Version 3.5
A platform that simulate rice crop growth based on inputs like weather, soil type, management practices.

Preparing the ORYZA Simulation inputs

• Weather data: All the weather data for this simulation is placed in this folder oryzatrain\test_data\ORYZA_weather_data\rach_gia. In ORYZA, weather data file format is organized by <station_name><station_code>.<year_of_weather_data>

  Here's what the weather file looks like in ORYZA format There are 9 columns in each weather file
  (1) Station number
  (2) Year
  (3) Day of year
  (4) Solar Radiation (KJ/m2/day)
  (5)(6)Min/Max Temperature (*C)
  (7) Vapor Pressure (kPa)
  (8) Wind Speed (m/s)
  (9) Rainfall (mm)

  At the first row of each weather file, it contains information of weather station: longitude; latitude; height, and 2 Angstrom parameters

• Soil file (standard.sol): This soil file contains soil properties of simulation area. In this simulation, I used soil file provided by Alexandre with the assumption that the place of the soil file has the same properties with simulation place since we haven't got the actual soil data yet

• Crop file (IR72.crp): This crop file contains rice type properties provided by IRRI

• Experimental file (standard.exp): This file lets you define and set up planting, irrigation and fertilization schedule for rice management. However, for this experiment, we do not run this file directly. Instead, we use rerun file (.rer) to control the specific irrigation treatments

• Rerun files (.rer): These are the main files you will run. They tell ORYZA to:

  1. Load the standard.exp file as a base template.
  2. Override the default water management section with a new, specific schedule

  This is how we compare CF and AWD. You can find these rerun files in a location like oryzatrain\scene1\CF_s1.rer or oryzatrain\scene1\AWD_s1.rer

  For Continuous Flooding (CF): We use a rerun file likeCF_s1.rer. In this file, the irrigation schedule is set to apply water frequently to keep the ponding water level high (e.g., irrigate when water is below 2 cm)

  IRMTAB = 0.0, 2.0,   ! Keep the plot flooding
          85.0, 0.0   ! Drain the plot before harvesting
  

  For Alternate Wetting & Drying (AWD): We use a rerun file like AWD_s1.rer. In this file, the irrigation schedule is delayed until the soil water potential reaches to threshold that equivalent to -15 cm

  IRMTAB = 0.0, 2.0,    ! Keep the plot flooding
          8.0, 3.0,    ! After 1 week switch to AWD
         60.0, 2.0,    ! Keep the plot flooding during flowering
         72.0, 3.0,    ! Switch to AWD
         85.0, 0.0     ! Drain the plot for harvesting
  

  In the both above code blocks, we useIRMTABas an irrigation option switching table and it has 2 columns

  IRMTAB = <day after emergence>, <irrigation option>,
      = <day after emergence>, <irrigation option>
  

  where the first column is value based on the day after rice planting, second column is irrigation option (SWITIR). In this case, each irrigation option (SWITIR) is defined in standard.erp file. For example, the ruled for second irrigation (2) is defined as following:

  IRRI2=50. ! Real Value
  WL0MIN=20. ! Real Value, irrigate when water is below 2 cm
  

  where IRRI2 is the amount of irrigation (mm) and WL0MIN is the minimum standing water (mm). And the ruled for third irrigation (3) is defined as following:

  IRRI3=50. ! Real Value
  KPAMIN=10. ! Real Value
  SLMIN3=2 ! Integer Value, Unit: KPa
  

  where IRRI3 is the amount of irrigation (mm) and KPAMINis critical soil water potential at layer SLMIN3 (AWD Threshold). SLIMIN3 is soil layer for critical value of KPAMIN.

Scenarios

You will run the CF and AWD simulations for each of these scenarios to compare their performance
1. Baseline: This uses the normal weather data and the standard AWD threshold
2. Temperature rises 2℃: This uses a rerun file that has been modified the weather module to raise the temperature to 2℃ (Apply for both CF and AWD)
3. AWD Threshold changing: This scenario only applies to the AWD simulation. The soil water potential has been modified to simulate the lower threshold (e.g., -20cm or -25cm) and let the field drain more

How to run the simulation

Select your scenario by choose the corresponding path. I've created 3 folders for 3 scenarios and each folder has 2 rerun files: CF, AWD.

• Scenario 1: \oryzatrain\scene1\CF_s1.rer \oryzatrain\scene1\AWD_s1.rer
• Scenario 2: \oryzatrain\scene2\CF_s2.rer \oryzatrain\scene2\AWD_s2.rer
• Scenario 3: \oryzatrain\scene3\CF_s3.rer \oryzatrain\scene3\AWD_s3.rer

For example, if you want to simulate CF practice when the temperature rises 2℃, you will set path of that scenario and practice control file at FILEIR in (CONTROL.DAT). This file used to control the simulation of ORYZA (Input/Output file path and option for output display)

	CONTROLFILE = 'CONTROL.DAT'
	FILEON = '.\res.dat'             ! Output file
	FILEOL = '.\model.log'           ! Log file
	FILEIT = '.\standard.exp'        ! Experimental data
	FILEI1 = '.\IR72.crp'       ! Crop data
	FILEIR = '.\scene2\CF_s2.rer'    ! Rerun file contains scenario and practice which you want to simulate
	FILEI2 = '.\standard.sol'        ! Soil data

And there is an executable file (ORYZA35.exe), you will run this program to execute the simulation by running this command inside the oryzatrain folder terminal

.\ORYZA35.exe

When the simulation is complete, it will generate 2 output files op.dat and res.dat.

1. res.dat: contains the detailed simulation results of the model in the daily time step
2. op.dat: contains the summary or end-of-simulation values (yields, crop duration, etc.)

Here's the list of variables in res.dat and op.dat files

res.dat

Variable name	Unit	Definition
TIME		Accumulative day number starting from Jullian day of simulation starting
DVS		Development stage number
TMIN	℃	Daily minimum temperature
TMAX	℃	Daily maximum temperature
LAI		Leaf area index
IR	mm/d	Amount of daily irrigation
WL0	mm	Depth of surface water
CO2C	Kg C/ha	Daily CO2 emission
CH4C	Kg C/ha	Daily CH4 emission
N2ON	Kg N/ha	Daily N2O emission
SOC	Kg C/ha	Daily organic carbon total in soil profile
SON	Kg N/ha	Daily soil organic nitrogen total in the soil profile
DOY		Julian day in a year
CROPSTA		Crop growth status: 1 for sowing, 2 for seed bed, 3 for transplanting, 4 for main field

op.dat

Variable name	Unit	Definition
RUNNUM		Number of runs in the simulation (depends on rerun set)
WRR14	kg/ha	Grain yield with 14% moisture
WAGT	kg/ha	Dry weight of above-ground plant organs
RAINCUM	mm	Accumulative rainfall in main field
IRCUM	mm	Accumulative irrigation in main field
SOC	kg C/ha	Soil organic carbon content at the end of season
SON	kg N/ha	Soil organic nitrogen content at the end of season
S_CH4C	kg C/ha	Seasonally accumulative CH4 emission
DAE		Days after emergence

For more information you can find the ORYZA manual in oryzatrain folder and you can check this sheet table which contains input data and parameters in ORYZA

GAMA

After running ORYZA simulation, I use GAMA to visualize the results from ORYZA and also as an input for the farmer decision making model.

ORYZA output files have.datformat so it is not easy for GAMA to process them. Hence, there are 2 python script files inpython-scriptdirectory used to convert them: process_op.py;process_res.py. You can modify the path of input file and output file as you want at the last code line

In process_op.py
df1 = convert_opdat_to_csv('<input path of op.dat>', '<output path of csv>')
In process_res.py
df1 = convert_res_to_csv('<input path of res.dat>', '<output path of csv>')

For the use of GAMA model, there are 3 GAMA files: seasonal_result.gaml,daily_result.gaml,eco_model.gaml

1. seasonal_result.gaml, This GAMA model will visualize the seasonal result (op file) for all 3 scenarios including Rice Yield; Irrigation; Methane Emission. There are 3 experiments corresponding to 3 indicators.
   Irrigation Experiment for water-use visualization

You can change the path of each op file as you want inaction load_data

action load_data {
	cf_s1 <- csv_file("../includes/Results/CF_s1/cf_op.csv");
	cf_s2 <- csv_file("../includes/Results/CF_s2/cf_op.csv");
	cf_s3 <- csv_file("../includes/Results/CF_s3/cf_op.csv");
	awd_s1 <- csv_file("../includes/Results/AWD_s1/awd_op.csv");
	awd_s2 <- csv_file("../includes/Results/AWD_s2/awd_op.csv");
	awd_s3 <- csv_file("../includes/Results/AWD_s3/awd_op.csv");
}

2. daily_result.gaml This GAMA model will visualize the daily result (res file) and you can choose 2 scenarios to compare the water level, Leaf Area Index (LAI), emission in a selected season together. There are 2 experiments: water_level, emission

   1. water_level
      This experiment contains plot map to visualize the water level for 2 practices. It also has 2 charts to show water level and LAI
      You can select Scenario and Season to compare the result (reload experiment is required). In Season selection, I still use the season index for selection
      
   2. emission
      This experiment contains 4 charts corresponding for 4 indicators that could be simulated in ORYZA: NO2; CO2; Soil Organic Carbon, CH4 You can select Scenario and Season to compare the result (reload experiment is required)

3. eco_model.gaml This GAMA model will use the seasonal results (yield simulation results) as an aspect for the farmer to make a practice decision after a season. However this model only compare 2 practices in a scenario, if you want to compare in other scenarios you have to change the path of op file in action load_data

   action load_data {
   	cf_s1 <- csv_file("../includes/Results/CF_s1/cf_op.csv");
   	awd_s1 <- csv_file("../includes/Results/AWD_s1/awd_op.csv");
   }
   

   So there are 10 rice seasons which are Winter-Spring and Summer-Autumn (double rice) from 2015 - 2020.

   At the beginning of the simulation (Winter-Spring 2015-2016), we initialize a number of farmers who use AWD practice (in this case is 40). Then after every season, CF farmers will compare their own income to the average income of nearby AWD farmers. If the AWD farmers' average income is higher by a certain threshold, the CF farmer will consider switching
   In this model, there are 2 species

   1. Plot
      Created and got info from the shapefile which has
      
      • land_use
      • area
      • plot_id
   2. Farmer
      Each farmer owns a plot and they can decide which practice they will apply on their plot (AWD or CF) For the economic aspect, each farmer has
      
      • Revenue
        
      • Cost
        
      • Profit
        Revenue of each farmer is calculated by the yield multiplies with farmer area and rice price.

      current_revenue <- current_yield * (my_plot.area) * rice_price; // yield(kg/ha)*area(ha)*price(VND/kg)
      

   And on the market, there are some factors relating to price/cost. These factors you can change during the simulation
   - Rice price
   - Fertilizer cost, seed cost, other cost
   
   The unit for economic factors is VND

   The simulation visual has 3 elements:
   

   • Plot window: Visualize the change of irrigation practice by farmer
     
   • AWD Adoption rate chart: Percentage of farmer switching to AWD
     
   • Income per ha: The change of farmer income per ha between 2 practices