Citation
Climate Change Impacts on Rice (Oryza Sativa) Yield and Adaptation Strategies for Rice Production in the Artibonite Valley of Haiti

Material Information

Title:
Climate Change Impacts on Rice (Oryza Sativa) Yield and Adaptation Strategies for Rice Production in the Artibonite Valley of Haiti
Creator:
Nicolas, Floyid
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (97 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Agricultural and Biological Engineering
Committee Chair:
Migliaccio,Kati White
Committee Co-Chair:
Hoogenboom,Gerrit
Committee Members:
Eisenstadt,William R
Rathinasabapathi,Balasubramani

Subjects

Subjects / Keywords:
adaptation-strategies -- artibonite-valley -- climate-change -- cropping-system-model -- food-security -- haiti -- rice-production
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Agricultural and Biological Engineering thesis, M.S.

Notes

Abstract:
Rice (Oryza sativa) is one of the major crops in the world and one of the most consumed agricultural products in Haiti. The Artibonite Valley is responsible for more than 70% of the rice production of the country. Rice production in Haiti has faced many threats, including weather uncertainty. The study objectives were to investigate the potential impacts of climate change on rice yield in the Artibonite Valley of Haiti for future periods (near-term: 2010-2039 and mid-century: 2040-2069) under two Representative Concentration Pathways (4.5 and 8.5). The DSSAT CERES-Rice model was used to perform the simulations using local soil, meteorological and crop experimental data including management practices, phenology and yield. The results indicated that the temperatures are predicted to increase in all three rice-growing seasons (spring-summer, summer-autumn and winter-spring). Under both RCPs (4.5 and 8.5), the simulation results indicated that ensemble-average rice yield decreased in the spring-summer and summer-autumn seasons (5.1-6.6% and 5.4-8.3%) and increased in the winter-spring season (2.3-3.6%). Adaptation strategies which include switching planting dates, breeding new cultivars and changing fertilizer application rate improved the average annual rice yield by 11-17%, 10-26% and 14-23%, respectively by the mid-century in the Artibonite Valley compared to the annual average yield decrease (5.2-7.0%). These findings are a significant contribution from an adaptation perspective for policymakers, farmers and stakeholders in rice production in Haiti. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2019.
Local:
Adviser: Migliaccio,Kati White.
Local:
Co-adviser: Hoogenboom,Gerrit.
Statement of Responsibility:
by Floyid Nicolas.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
LD1780 2019 ( lcc )

UFDC Membership

Aggregations:
University of Florida Theses & Dissertations

Downloads

This item has the following downloads:


Full Text

PAGE 1

CLIMATE CHANGE IMPACTS ON RICE ( Oryza sativa ) YIELD AND ADAPTATION STRATEGIES FOR RICE PRODUCTION IN THE ARTIBONITE VALLEY OF HAITI By FLOYID NICOLAS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2019

PAGE 2

2019 Floyid Nicolas

PAGE 3

To the loving memory of my father, Phazil Nicolas. To my mother Fernande Coffy and my siblings, Magarette, Naomie, Bath Sheba, Tamara and Peter Nicolas

PAGE 4

4 ACKNOWLEDGMENTS The process of writing this thesis was fascinating and rewarding. I want to thank all the people who have contributed to the realization of this study in many different ways. First, I would like to express my gratitude to my supervisor Dr. Kati W. Migliaccio for her incredible patience, her en couragement, her useful remarks comments and engagement through out the learning process of this master thesis. Professor Migliaccio office was always available whenever I ran into trouble or had a question about my research. H er high level of efficiency h as made a tremendous impact on me. Furthermore, I would like to thank all my other committee members whose continuous counsel allowed this project to be successfully done. I want to thank Dr. Gerrit Hoogenboom for his critical remarks, suggestions, and bei ng a role model; Dr. William R Eisenstadt for his continuous encouragement his help with the onsite weather data; Dr. Bala Rathinasabapathi for his support, comments, and encouragement. I also want to thank the AREA project and the USAID that funded this p roject and provide me with this opportunity to realize this master. Many thanks to Dr. Rose Koenig for her unconditional support. Thanks also to Dr. Absalon, Dr. Caroline Staub, and Mrs. Nicole Monvale, Mr. Maurice W iener, and Mrs. Sabine Vales for their s upport. Finally, I want to thank Mr. Mid r ouin Lidelias for helping me to have access to the rice data from the Ministry of Agriculture of Haiti, Mr. Gary Doliscar for his useful advice and his help in getting some field data in Haiti that were essential t o this study.

PAGE 5

5 TABLE OF CONTENTS P age ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 7 LIST OF FIGURES ................................ ................................ ................................ ......................... 8 LIST OF ABBREVIATIONS ................................ ................................ ................................ ........ 10 ABSTRACT ................................ ................................ ................................ ................................ ... 12 1 LITERATURE REVIEW AND RATIONAL ................................ ................................ ........ 14 Climate Change ................................ ................................ ................................ ...................... 14 Climate Change, Agriculture, and Food Security ................................ ................................ ... 15 Climate Change Impacts in Haiti ................................ ................................ ............................ 16 Methods for Estimating Climate Change Impacts on Agricultural Production ...................... 17 Adaptation Strategies to Climate Change ................................ ................................ ............... 21 Agriculture and Rice Production in Haiti ................................ ................................ ............... 22 Problem Statement ................................ ................................ ................................ .................. 23 Objectives ................................ ................................ ................................ ............................... 25 2 ASSESSING THE POTENTIAL IMPACT OF CLIMATE CHANGE ON RICE YIELD IN THE ARTIBONITE VALLEY OF HAITI USING THE CSM CERES RICE SIMULATION MODEL ................................ ................................ .............................. 28 Introduction ................................ ................................ ................................ ............................. 28 Materials and Methods ................................ ................................ ................................ ........... 31 Study Site ................................ ................................ ................................ ......................... 31 Weather Data for Model Input ................................ ................................ ........................ 32 Experimentally Collected Data for the Crop ................................ ................................ ... 33 Soil Data ................................ ................................ ................................ .......................... 34 DSSAT CERES Rice ................................ ................................ ................................ ...... 35 Calibration and Evaluation of th e Model ................................ ................................ ........ 35 Climate Data ................................ ................................ ................................ .................... 37 Model Scenario Simulations ................................ ................................ ............................ 38 Results ................................ ................................ ................................ ................................ ..... 39 Calibration and Evaluation of the Model ................................ ................................ ........ 39 Future Climate Scenarios ................................ ................................ ................................ 40 Simulation of Climate Change Impacts on Rice Production ................................ ........... 42 Changing climate effects on phenology of TCS10 ................................ .................. 43 Changing climate effects on phenology of CAP ................................ ...................... 43 Changing climate effects on rice yield ................................ ................................ ..... 44 Discussion ................................ ................................ ................................ ............................... 45

PAGE 6

6 Model Performance ................................ ................................ ................................ ......... 45 Model Simulation ................................ ................................ ................................ ............ 46 Conclusion ................................ ................................ ................................ .............................. 48 3 SIMULATING POTENTIAL ADAPTATION STRATEGIES TO ALLEVIA TE CLIMATE CHANGE IMPACTS ON RICE YIELD IN THE ARTIBONITE VALLEY OF HAITI ................................ ................................ ................................ ............................... 61 Introduction ................................ ................................ ................................ ............................. 61 Materials and Methods ................................ ................................ ................................ ........... 63 Experimental Site ................................ ................................ ................................ ............ 63 DSSAT Cropping S ystem Model ................................ ................................ .................... 64 Weather, Experimental Rice and Soil Data and Model Performance ............................. 65 Modeling Simulations ................................ ................................ ................................ ..... 66 Assessment of Potential Agronomic Adaptation Measures ................................ ............ 66 Changing the transplanting date ................................ ................................ ............... 67 Breeding new rice cultivars ................................ ................................ ...................... 67 Altering Fertilizer Application Rate ................................ ................................ ......... 68 Results ................................ ................................ ................................ ................................ ..... 68 Change in Transplanting Dates ................................ ................................ ....................... 68 Breeding New Rice Varieties ................................ ................................ .......................... 69 Altering Fertilizer Application Rate ................................ ................................ ................ 70 Discussion ................................ ................................ ................................ ............................... 72 Conclusion ................................ ................................ ................................ .............................. 74 4 SUMMARY OF FINDINGS AND CONCLUSION ................................ ............................. 80 Chapter 1 ................................ ................................ ................................ ......................... 80 Chapter 2 ................................ ................................ ................................ ......................... 81 Chapter 3 ................................ ................................ ................................ ......................... 82 LIST OF REFERENCES ................................ ................................ ................................ ............... 84 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ......... 97

PAGE 7

7 LIST OF TABLES Table page 2 1 Input parameters of management practices of the TCS10 cultivar for the DSSATCERES Rice model. These data are from field experiments conducted in the Artibonite Valley of Haiti by th e Ministry of Agriculture of Haiti. ................................ .. 50 2 2 Input factors of management practices of the CAP variety for the DSSATCERES Rice model. These data are from field experiments conducted in the Artibonite Valley of Haiti by the Ministry of Agriculture of Haiti. ................................ .................... 51 2 3 Soil data of the Mauge farm in the Artibonite Valley (Haiti). ................................ ........... 52 2 4 Genetic coefficients for rice cultivar and their definitions in the DSSAT CERES rice model (Hoogenboom et al., 2017). ................................ ................................ .................... 52 2 5 Information related to the five GCMs used in this rese arch study. ................................ ... 53 2 6 Current planting and harvesting dates of the farmers of the Artibonite Valley. These planting dates were simulat ed in the DSSAT seasonal tool to evaluate the impacts of the climate change on rice yield for each growing rice seasons. ................................ ....... 53 2 7 Ge netic coefficient values (unitless) for TCS10 and CAP rice cultivar after calibration with GLUE. These two varieties are among the widely cultivated rice cultivars in the Artibonite Valley of Haiti. ................................ ................................ ........ 53 2 8 Calibration and evaluation of the cultivars (TCS10 and CAP) in Artibonite Valley in Haiti. ................................ ................................ ................................ ................................ ... 54 3 1 Grain yield (kg ha ) of virtual rice cultivars derived from TCS10 under baseline climate and projected changes in temperature, CO 2 and rainfall by for the near term and the mid century climate periods at the Artibonite Valley, Haiti. ................................ 77

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Emissions of carbon dioxide (CO 2 ) alone in the Representative Concentration Pathways (RCPs) (lines) and the associated scenario categories (colored areas show 5 to 95% range). Copi ed from IPCC (2014) ................................ ................................ ...... 26 1 2 Climate change scenarios. (a) Change in annual mean surface temperature, (b) change in annual mean preci pitation, in percentages, and (c) change in average sea level by Coupled Model Intercomparison Project Phase 5 (CMIP5) multi model mean projections (i.e., the average of the model projections available) for the 2081 2100 period under the RCP2.6 (left) and RCP8.5 (right) scenarios. Copied from IPCC (2014). ................................ ................................ ................................ ...................... 27 2 1 Region of rice production in the Artibonite Valley (MARNDR, 2013) ............................ 49 2 2 Relationship of observed grain yield with simulated yield of rice for (a) TCS10 and (b) CAP cultivars used in the study. These observed data were collected in experimental trials conducted in the Artibonite Valley of Haiti in 2012, 2014, 2015 and 2016 by the Ministry of Agriculture of Haiti. These data were not used in the calibration process. The dashed line represents the line 1:1 and the red line is the regression line. ................................ ................................ ................................ ................... 55 2 3 Predicted variation of the maximum and minimum temperatures with five GCMs during the three seasons (spring summer, summer autumn and winter spring) for two future climate periods (near term:2010 2039; and mid century: 2040 2069) under both RCPs 4.5 and 8.5 compared with the baseline. Each of the periods is 30 years of climate d ata. The fives GCMs are on the top x axis and the three seasons are on the right y axis. The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively. ................................ ................................ .. 56 2 4 Predicted variation of the solar radiation and rainfall with GCMs in the three seasons (spring summer, summer autumn and winter spring) for two future climate periods (near term:2010 2039; and mid century: 2040 2069) under both RCPs 4.5 and 8.5 compared with the baseline. Each of the periods is 30 years of climate data. The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, r espectively. ................................ ................................ ............................. 57 2 5 Predicted changes in the flowering duration (a) and maturity duration(b) for two future climate periods (near term and mid century) under two RCPs (4.5 and 8.5) during three seasons (spring summer, summer autumn and winter spring). The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively. The flowering a nd maturity days for each of the three rice growing seasons are on the left y axis, the cultivars are on the bottom x axis and the GCMs are on the top y axis. ................................ ................................ ........................ 58

PAGE 9

9 2 6 Simulated yield (kg/ha) of the TCS10 and CAP rice cultivars for two climate periods (near term and mid century) during three seasons (Spring summer, summe r autumn and winter spring) with the five GCMs in the Artibonite Valley under the RCP 4.5 compared to the baseline. The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively. The cultivars (TCS10 and CAP) are in the bottom x axis, the GCMs are on the top x axis and the yield for each of the tree rice growing seasons are on the left y axis. ................................ .............. 59 2 7 Simulated yield (kg/ha) of the TCS10 and CAP rice cultivars for two climate periods (near term and mid century) during three seasons (Spring summer, summer autumn and winter spring) with the five GCMs in the Artibonite Valley under th e RCP 8.5 compared to the baseline. The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively. The cultivars (TCS10 and CAP) are in the bottom x axis, the GCMs are on the top x axis and the yield for each of the tree rice growing seasons are on the left y axis. ................................ .............. 60 3 1 Rice transplanting dates for the spring summer, summer aut umn and winter spring seasons in the Artibonite Valley under the RCP 4.5 of climate change. The weeks are in the top axis of the graph and the two rice cultivars are in the bottom axis. The green, red and dark red colors represent the baseline, near term a nd mid century climate periods. The dark red boxes are the current planting dates for the spring summer, the summer autumn and the winter spring seasons, respectively. ...................... 75 3 2 Rice transplanting dates for the spring summer, summer autumn and winter spring seasons in the Artibonite Valley under the RCP 8.5 of climate change. The weeks are in the top axis of the graph and the two rice cultivars are in the bottom axis. The green, red and dark red colors represent the baseline, near term and mid century climate periods. The dark red boxes are the current planting dates for the spring summer, the summer autumn and the winter spring se asons, respectively. ...................... 76 3 3 Adaptation strategies for change in fertilizer rate application (50, 100,130, 150, 180, 210, and 250 k g N ha 1 ) for the near term (2010 2039) and the mid century (2040 2069) under the RCP 4.5 climate scenario. ................................ ................................ ........ 78 3 4 Adaptation strategies for change in fertilizer dose application (50, 100,130, 150, 180, 210, and 250 kg N ha 1 ) for the near term (2010 2039) and the mid century (2040 2069) under the RCP 8.5 climate scenario. ................................ ................................ ........ 79

PAGE 10

10 LIST OF ABBREVIATIONS AgMIP Agricultural Model Intercomparison and Improvement Project APSIM Agricultural Production Systems Simulator AT After transp l antation CAP Crte Pierrot CEC Electrical conductivity CEC Cation exchange capacity CMIP5 Coupled Model Intercomparison Project Phase 5 CNSA Coordination National de Security Alimentaire CO 2 Carbone d ioxide CROPWAT A computer program for irrigation planning and man a gement CSM Cropping System Model DSSAT Decision Support System for Agrotechnology Transfer EPIC Environmental Policy Integrated Climate GCM Global C limate M odel GLUE Generalized Likelihood Uncertainty Estimation INRA Institut National de la Recherche Agronomique K Potassium MACROS Modules of an Annual CROp Simulator MARNDR Rural N Nitrogen

PAGE 11

11 NASA POWER National Aeronautics and Space Administration P r ediction of Worldwide Energy Resources NRMSE Normal Root Mean Square Error OC Organic Carbon OM Org anic Matter ORYZA1 Ecophysiological Model for Irrigated Rice Production P Phosphorus PRECIS Providing Regional Climates for Impacts Studies RCM Regional Climate Model RMSE Root Mean Square Error STICS Simulateur mulTIdiscplinaire pour les Cultures Standard TCS10 Taichung Sen 10 USAID United States Agency for International Development

PAGE 12

12 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science CLIMATE CHANGE IMPACTS ON RICE ( Oryza sativa ) YIELD AND ADAPTATION STRATEGIES FOR RICE PRODUCTION IN THE ARTIBONITE VALLEY OF HAITI By Floyid Nicolas August 2019 Chair: Kati W. Migliaccio Major: Agricultural and Biological Engineering Rice ( Oryza sativa ) is one of the major crops in the world and one of the most consumed agricultural products in Haiti. The Artibonite Valley is responsible for more than 70% of the rice production of the country Rice production in Haiti has faced many threats, including weather uncertainty The study objectives were to investigate the po tential impacts of climate change on rice yield in the Artibonite Valley of Haiti for future periods (near term: 2010 2039 and mid century: 2040 2069) under two Representative Concentration Pathways (4.5 and 8.5). The DSSAT CERES Rice model was used to perform the simulations using local s oil, meteorological and crop experimental data including management practices, phenolog y and yield. The results indicated that the temperatures are predicted to increase in all three rice growing seasons (spring summer, summer autumn and winter spring). Un der both RCPs (4.5 and 8.5), the simulation results indicated that ensemble average rice yield decrease d in the spring summer and summer autumn seasons (5.1 6 .6% and 5.4 8.3%) and increase d in the winter spring season (2.3 3.6% ). Adaptation strategies whic h include switching planting dates, breeding new cultivars and changing fertilizer application rate improved the average annual rice yield by 1 1 17%, 10 2 6 % and 1 4 23 %, respectively by the mid century in the Artibonite Valley compared to the annual

PAGE 13

13 average yield decrease (5. 2 7.0 %). These findings are a significant contribution from an adaptation perspective for policymakers, farmers and stakeholders in rice production in Haiti.

PAGE 14

14 CHAPTER 1 LITERATURE REVIEW AND RATIONAL E Climate C hange Climate change denotes the variation in the long term weather patterns in the regions of the world due to anthropogenic actions. These acts have led to an elevation of atmospheric concentrations of greenhouse gases (GHG) which in clude carbon dioxide (CO 2 ), nitrous oxide ( N 2 O ) and methane ( CH 4 ) as never recorded in the last 800,000 years. Climate change has been observed since 1950 as the atmosphere has warmed, the sea level has risen, and the amount of snow and ice have diminished (Pachauri et al., 2014) The factors which drive the anthropogenic greenhouse gas emissions are mostly economic activities, population dynami cs, lifestyle, energy use, land use changes, technology and climate policy. The Representative Concentration Pathways (RCPs) project identified four different scenarios of GHG emissions and air pollutant emissions, atmospheric concentrations and land use b y the 21st century. The four climate scenarios are RCPs 2.6, 4.5, 6.0, and 8.5 compar ed to a reference (observed climate information). The RCP 2.6 represents a rigorous mitigation scenario by assuming constant emissions after 2100 aiming to limit the eleva tion of global mean temperature to 2 C above pre industrial temperatures (Van Vuuren et al., 2011) The RCP 6.0 predicts a stabilization of radiation forcing at 6.0 Wm 2 by the end of the century (Masui et al., 2011) The RCP 4.5 assumes a stabilizati on of the emissions at the end of the century (Thomson et al., 2011) while the 8.5 considers a considerable increase of greenhouse gas emission, leading to a radiative forcing of 8.5 Wm 2 by 2100 (Riahi et al., 2011) The fifth assessment report (AR5) of the Intergovernmental Panel on Climate Change (IPCC, 2014) provided strong evidence for climate change, which includes an increase of the global mean surface temperature up to 4 C and rise of the global mean sea level up to 0.8 m.

PAGE 15

15 The mean precipit ation is predicted to decrease in the mid latitude and subtropical regions and increase in the mid latitude wet region. These predictions were made based on the Coupled Model Intercomparison Project Phase 5 (CMIP5) ensemble under the most extreme climate s cenario (RCP 8.5) (Figures 1 1 and 1 2). Climate C hange, A griculture, and F ood S ecurity The changes in climate factors such as temperature, precipitation, and extreme weather events are predicted to affect agriculture and food security across all regions of the world, particularly in developing countries where most of the population relies on agriculture and natural resources (Nath and Mandal, 2018; Wunder et al., 2018) Both climate variability and climate change affect weather patterns, as well as the frequency and harshness of extreme climate events. Agricultural production and natural ec osystems will be profoundly affected in the areas which are already vulnerable to food insecurity, with the largest impacts being decreased crop yields and the productivity of the livestock (Fanzo et al., 2018) Matthews and Wassmann (2003) stated that the increase in carbon dioxide, temperature and uncertainty in rainfall due to climate change could have a considerable effect on crop growth, development, and yield. Also, they highlighted the pressure of the increasing population on the food production of the world. In addition to factors limiting global agricultural production, weather patterns represent another threat on the system of the are projected to trigger yield increase at higher latitudes and yield decrease at lower lat itudes (IPCC, 2014) Haile et al. (2017) estimated the possible effects of climate change on worldwide food production and found that global crop production is predicted to be reduced by up to 9% in the 2030s and 23% in 2050s on average with enormous h eterogeneity across both the crops and countries. Warming temperatures and higher intensity of weather events due to climate change during the plant growing periods are predicted to impact negatively crop production by

PAGE 16

16 decreasing the yield of the major cultivated crop across the world (Ray et al. 2015; Rosenzweig et al. 2001) Climate C hange I mpacts in Haiti Many studies have shown the potential of the intensification of the occurrence and severity of extreme weather events, such as hurricane s in the poorest countries in the world. Many aspects of human rights such as poor education and health facilities in these countries have led to the exposure of people to high levels of risk and vulnerability (Mal et al., 2018 ) Although the relationsh ip between cyclones and of climate change remains elusive (Walsh et al., 2016) models have found an amplified severity of tropical storms in regions with warmer climate across the world (Patricola and Wehner, 2018; Van Oldenborgh et al., 2017) Haiti sits and is hit regularly during hurricane season by tropical storms that devastated hurricanes in the past 30 years, including Hurr icane Jeanne, which killed thousands in Gonaives in 2004 (USAID, 2017) In 2008, four storms (Ike, Fay, Hanna, and Gustav) struck the country and destroyed more than 60% of crops and triggered the death of more than 3,000 people (Cohen and Singh, 2014) More recently, Hurricane Matthew str uck the Southwestern part of the country in 2016, leaving 1.5 million in need of humanitarian relief and affected crop production significantly R ice yields are expected ons due to the change of the precipitation patterns, higher temperatures, and other climate related factors (Welch et al., 2010) Cohen and Singh (2014) re ported that radical changes had been observed in climate such as in the seasonality of rainfall, the temperature and the frequency and intensity of the tropical storms, which has triggered flooding and erosion The Haitian Ministry

PAGE 17

17 R essources Naturelles, et de Dveloppement Rural, or MARNDR) ha s observed temperature variation from 1973 to 2003 and reported that the average temperature rose by more than 1 o C during that period Haiti was ranked as the most vulnerable country to the chan ging climate in the world according to the climate change risk index (Maplecroft, 2016) The rural regions are the most sensitive to these changes. Between 2010 and 2014, crop deficits occurred in 65% of the municipalities and these losses were associate d with climate variability (Tiepolo and Bacci, 2017) For the Caribbean region where Haiti is located the projections of the IPCC (2014) predict an approximate increase of 1.2 to 2.3 C in surface temperature, about 5 % decrease in precipitations and between 0.5 0.6 m sea level rise by the year 2100 compared to the baseline 1986 2005 In case these changes happen as predicted, the impact o n the availability of water resources and crop production would be substantial (Borde et al., 2015) Methods for E stimating C limate C hange I mpacts on A gricultural P roduction Different methods and approaches have been used for estimating the impact of cli mate change on crop production. Mendelsohn and Massetti (2017) highlighted four main methods to measure climate change effects on agriculture: controlled laboratory experiments, Ricardian/cross sectional climate studies, panel weather studies and agrono mic crop models. The four methods each have strengths and limitations in application to climate change research. Artificial controlled conditions in growth chambers have been used in laboratory experiments to grow crops to isolate the influence of the inc rease in both temperature and CO 2 on crop yields (Adams et al., 2001; Jena et al., 2018) The cross sectional method has been used adaptations measures (Sark er et al., 2014) However, this method does not assess the impacts for a specific crop or livestock (Mendelsohn and Massetti, 2017) and cannot evaluate climate change

PAGE 18

18 on yield variability (Sarker et al., 2014) The panel data approach allows overcoming the issues of assessing climate change impacts on a particular crop and omitted variables by allowing the use of fixed effects. It also has more degrees of freedom which facilitate the estimate of non linear effects of climate change (Blanc and Schlenker, 2017) Panel approaches using weather variation instead of climate variation have shown some limitations about omitting climatic variables (wind speed and relative humidity) other than temperature and precipitation. Theses omissions tend to bias the projected climate change impacts on the yield results (Zhang et al., 2017) The fourth method, agronomic models, are biophysical crop simulation models that have been used to predict crop growth. They are dynamic systems models that predict cro p yield, biomass, and changes in the biophysical outcomes, viz., soil moisture and soil carbon, by linking input factors (crop genetic coefficients soil characteristics, water dynamics nutrients, weather information, management practices) to the crop gro wth stages such as emergence, anthesis, panicle initiation and physiological maturity (Antle and Stckle, 2017) Jones et al. (2017) highlighted many different existing crop simulation models with their capabilities and limitations in simulating the co ntinual changes of the agricultural systems which include climate change. Different cropping system models including EPIC, APSIM, STICS and DSSAT have been used for long term simulation experiments. The Environmental Policy Integrated Climate (EPIC) mod el have been applied to simulate crop yield, soil water dynamics, and nutrient loss through runoff and leaching (Izaurralde et al., 2006; Worou et al., 2012) The Agricultural Production Systems Simulator (APSIM) model has been used to simulate crop yiel ds with climate trends and irrigation management (Asseng et al., 1998; Dixit et al., 2018) STICS ( Simulateur mulTIdiscplinaire pour les Cultures Standard ) is a crop model that has been developed at the Institut National de la Recherche Agronomique ( INRA ) (Brisson et al., 1998) to

PAGE 19

19 simulat e crop production and environmental impacts of cropping systems. This model has been used to simulate many agro ecosystems response management practices and environmental conditions at many different spatial and temporal scales systems (Constantin et al., 2012; Coucheney et al., 2015) The Decision Support System for Agrotechnology Transfer (DSSAT) (Hoogenboom et al., 2017; Jones et al., 1998; Jones et al., 2003; Tsuji, 1998) was developed to allow the application of crop models in a systems approach to the field of agronomic research. Many options such as crop phenotype, soil effects, weather, climate, and management decisions were added in the different DSSAT improved versio n (4.7). Crop models utilization has considerably increased in modern agricultural research, and they are recognized as useful tools. They have been helpful to many different research purposes as well as assessing the impacts of climate change on crop and predicting yields (Jones et al., 2003) Crop models are considered one of the consistent tools for the evaluation of the potential effects of climate change on crops and the efficacy of the adaptations measures (Jones et al., 2003; Van Ittersum et al., 2013) Matthews et al. (1997) concluded that ORYZA1 was able to simulate the likely changes in rice yield that could be triggered by variations in temperature and atmospheric concentration CO 2 but predicted CH 4 emissions poorly (Olszyk et al., 1999) CERES Rice model, on the other hand, i s capable of evaluating the CH 4 emissions, and the potential variation of the rice yield because it includes routines of soil organic matter decomposition routing and other tools that describe the appropriate crop ma nagement. The CERES R ice, EPIC and ORYZA2000 are reported as the most commonly used models (White et al., 2011) However, CERES R ice has been well tested in various type s of environments and ha s been able to sufficiently simulate the

PAGE 20

20 rice growth and deve lopment under a range of conditions which include upland and lowland (Vaghefi et al., 2013) The simulation of climate change impacts o n crop yield carries uncertainties due to variation among the crop models as in the downscaled global circulation mode ls. Therefore the use of multimodel s is suggested to produce more reliable results by improving the relationship between CO 2 and temperature (Asseng et al., 2013) The Agricultural Model Intercomparison and Improvement Project (AgMIP) (Rosenzweig et al., 2018) has developed new methods for Coordinated Global and Regional Assessments (CGRA) of agriculture and food security under the changing climate. These CGRA co nsistently link models scales and disciplines to address the complex chain of climate impacts and detect the main weaknesses feedbacks and uncertainties in handling future risk In t he AgMIP Rice ( http://agmip.org/r ice/ ) 13 rice models were evaluated different sites with varied climatic conditions in Asia and their contrasting modeling approaches on key physiological processes Individual crop model s trigger much more uncertainties compared to mu l tiple crop model s However, the average of simulations of all crop models carries an uncertainty of less than 10% of observed yields when reproducing experimental data (Li et al., 2015) Although significant advances have been made in the improvement of the quality of the agricultural system data, the knowledge systems, and the models, inherent uncertainties in the model structures are limitations that will persist and differ with the applications. Depending o n the applications, the agricultural system models are considered sufficient nonetheless significant advances are required to address the more challenging issues related to the climate trend and tackle food insecurity throughout the next century (Jones et al., 2017)

PAGE 21

21 Adaptation S trategies to C limate C hange Many studies have used DSSAT with a regional or global climate model to simulate the effects of climate on rice yield on national and field scales (Chun et al., 2016; Jayanta et al., 2010; Lychuk et al. 2017; Pranuthi and Tripathi, 2018) Instead of working at a plot scale, Xiong et al. (2009) coupled the same model with a regional climate model (RCM) to evaluate the impacts of both climate variability and climate change and the effects of CO 2 fertilization on 2 fertilization effects increased the rice yield but were not sufficient to offset the negatives impacts of the increased temperature. The CSM C ERES R ice has been used to forecast maturity and yield regarding both weather variability and chang ing climate, and has been tested and validated in many parts of the world including America, Europe, Africa, and Asia, (Dass et al., 2012; Kontgis et al., 2019; Singh et al., 2014; Singh et al., 2016) The simulation of adaptation strategies is one of the means that could provide an insight to the impacts of climate change in the future and it has been applied and gained recognition at both regional and i nternational levels (Porter et al., 2014; Tao and Zhang, 2010) The IPCC (2014a) highlighted and focused on various adaptive strategies such as cultivar breeding, irrigation, fertilization optimization, and planting date adjustment to mitigate the pote ntial effects of the chang ing climate on crop productivity (Xu et al., 2017) Porter et al. (2014) stated that the constant emphasis of the IPCC on these adaptation options is a practical decision to mitigate the impacts of climate change. Even though these adaptive measures have been simulated at local, regional (Babel et al., 2011; Gruda et al., 2019 ; Kim et al., 2013a; Moore and Lobell, 2014) and global scales (Challinor et al., 2014; Duong et al., 2019) application of the adaptive measures effectively remains a challenge.

PAGE 22

22 Agriculture and R ice P roduction in Haiti The agricultural sector is a critical socio ecological system in Haiti. Haitian agriculture has the highest employment rate in the entire Caribbean region with the involvement of nearly 60% of the active population of the country (Bargout and Raizada, 2013) Haitian agriculture is es sentially dedicated to self subsistence and is composed mostly of small farms. Current agricultural production in Haiti struggles to meet the national food demand. This is explained by a relatively low productivity level resulting from factors that include soils degradation due to inappropriate and intensive agricultural practices, lack of means of production, lack of improved seeds, and mismanagement of water resources (Cochrane et al 2016) Rice is one of the most consumed agricultural products in Ha iti after replacing traditional crops such as corn, tubers, and millet for Agricultural Organization (Food and Agriculture Organization of the United Nations, (FAO), 2013) reported that rice supply per person exceeded 50 k g and accounted for 21 percent of the mean of total calories consumed per day. The r ice consumption in Haiti started to increase in 1986 when Haiti opened its market to imported rice. Prior to that time, the country was self sufficient with rice (Cochrane et al., 2016) T he annual national consumption was estimated to be 450,000 tons of husked rice in 2014 c omp ared to 171,000 tons in 1985 (CNSA, 2014) From the early 1960s to 1980s, Haiti was self sufficient in rice T he average yield was higher at 2.5 ton/ ha; however, after the 1980s rice production declined. Haiti rice yield has averaged 1.83 ton/ha since 2 005 from an average of 2.20 ton/ha in the previous decade. The deterioration of land quality, failures to maintain and improve the irrigation infrastructure, inadequate water management, and lack of access to inputs are considered as the main causes of low er rice yields. Rice, one of the leading agricultural commodities that are widely used in rations in Haiti, is not sufficient to feed the country and the production has not increased over

PAGE 23

23 time (Cochrane et al., 2016) Among the six varieties of rice such as Malaka, Bogapot, Tididi, Schela, CAP and TCS10 are grown in the Artibonite Valley of Haiti. TCS10 and CAP are produced for the ir yield and Shela for its culinary quality (Lamy et al., 2011) The TCS10 variety belongs to the species Oryza sativa L. and the subspecies indica and is originally from Taiwan. It is one of the most popular with more than 60% of the market share. TCS10 is a short term production cultivar with a growing season of 4 months with short gra ins and has a relatively higher yield. Farmers prefer TCS10 cultivar due to its resistance to biotic threats such as empty head syndrome and spider mite. The CAP cultivar native to Haiti, and it is one of the most appreciated varieties in Haiti because of its yield and quality. CAP rice variety was threatened by "black straw" disease but remained in the Valley because of its market values (Rgis, 2016) Problem S tatement Among the obstacles hampering rice production in Haiti, climate change remains a challenging one as the weather is an essential factor influencing crop growth Based on the sensitivity of crop yield to the weather estimating the effects of climate trends on rice production systems in Haiti is relevant to the development of adequate strategies to adjust and mitigate the possible outcome for increasing sustainable productivity. The national average yield of paddy rice was estimated to 2.5 ton/ha in 2017, which is low compared to other areas w here rice is producing in the Latin and Caribbean region such as the Dominican Republic (6.5 ton/ha) and Colombia (5.9 ton/ha) (FEWS NET, 2018) The already low yield is challenged by weather variability which increase s the level of vulnerability of the country to food insecurity (Borde et al., 2015) Therefore, determining how the changing climate will impact rice yield in the Artibonite Valley is relevant in terms of regional and national food security.

PAGE 24

24 Flooding, intense rainfall, landslides, hurricanes, drought, soil erosion, and saltwater intrusion are the climate related events that will likely affect many aspects of the economy and the society in Haiti. The main sectors that are predicted to suffer from these climate impacts are agricultural production, water resources, and rural and coastal communities. The more intense storms coupled with rainfall reduction will likely impact the agricultural production negatively by decreasing the staple crop yield such as rice, corn, and pota toes, and worsen the vulnerability to food insecurity (USAID, 2013) second national communication on climate change report (2013) predicted an increase of t he maximum temperature up to 1.7 o C and the minimum temperatures up to 1.3 o C by 2070 compared to a baseline of observed data from 1971 to 2000 The rainfall was predicted to decrease up to 19% and the rice water requirements were predicted to increase by 7% by 2070 due to the effects of cl imate changes. Due to the inexistence of sufficient data, the study did not evaluate the potential impacts of the changing climate on crop yield Hence, studies that use well calibrated biophysical models with experimental crop and local soil data are necessary to assess the possible impacts of climate change on agricultural crop production in Haiti in order to help with the decisions making of th e stakeholders including policy makers NGOs and farmers. CERES rice has been used in many regions in the world (White et al., 2011) However, the application of the model was not found in the literature for the Caribbean countries including Haiti. C lim ate change severely threatens the already recorded low rice yield of Haiti. Given the importance of the rice production in Haiti and its vulnerability to various factors including climate, this widely used crop model was used to simulate climate scenarios impacts that are conducive to decision making in term of mitigation of the global warming effects on rice productivity and open the path to other climate studies in the agricultural field of the country

PAGE 25

25 Objectives The overall objective was to simulate the climate change impacts on rice yield in the Artibonite Valley (Haiti) and evaluate adaptations strategies to improve to the rice cultivation under the changing climate The specific objectives include d to: Determine the climate change impacts on rice yield in Artibonite Valley of Haiti under both the RCP 4.5 and 8.5 for the near term ( 2010 2039) and the mid century ( 2040 2069 ) as compared to a baseline of 30 years. Evaluate adaptation strategies to identify rice production practices that rever se the potential negatives effects of the changing climate on rice production in the Artibonite Valley of Haiti during two future climate periods (2010 2039; 2040 2069) considering the two RCPs 4.5 and 8.5.

PAGE 26

26 Figure 1 1 Emissions of carbon dioxide (CO 2 ) alone in the Representative Concentration Pathways (RCPs) (lines) and the associated scenario categories (colored areas show 5 to 95% range). Copied from IPCC (2014)

PAGE 27

27 Figure 1 2 Climate change scenarios. (a) Change in annual mean surface temperature, (b) change in annual mean precipitation, in percentages, and (c) change in average sea level by Coupled Model Intercomparison Project Phase 5 (CMIP5) multi model mean projections (i.e ., the average of the model projections available) for the 2081 2100 period under the RCP2.6 (left) and RCP8.5 (right) scenarios. Copied from IPCC (2014).

PAGE 28

28 CHAPTER 2 ASSESSING THE POTENTIAL IMPACT OF CLIMATE CHANGE ON RICE YIELD IN THE ARTIBONITE VALLE Y OF HAITI USING THE CSM CERES RICE SIMULATION MODEL Introduction The increase of greenhouse gases (GHG) such as carbon dioxide (CO 2 ), methane (CH 4 ) and nitrous oxide (N 2 O) due to anthropogenic actions has led to climate change with worldwide economic and environmental implications. Climate change has impacted the cyclic pattern of the weather conditions and has triggered an increase in temperature, an increas e in the frequency of extreme weather events and a decrease in precipitation and for many locations across the globe (Pachauri et al., 2014; Pranuthi and Tripathi, 2018) The Intergovernmental Panel on Climate Change (IPCC) in its fifth assessment report (AR5) states that precipitation w ill likely chan ge depending on the region and the global mean temperature could rise by 4.8 o C by the end of the 21st century (Pachauri et al., 2014) Both the social and natural systems are already affected by climate change and increasing weather fluctuation will af fect the livelihoods of those near poverty who heavily depend on natural resources. T he permanent alteration of the ecosystems may affect water availability (Ding et al., 2017; Wunder et al., 2018) Also, extreme weather events like heavy rains, storms a nd droughts could considerably reduce natural resources, increase the soil erosion process and increase the risk of crop harvest failures (IPCC, 2014) Extreme weather events coupled with temperature increase and precipitation decrease are predicted to a ffect negatively on agriculture productivity as crop production systems are sensitive to the changes in the climate parameters. The increase of the mean seasonal temperature could lead to the reduction of the crop growth stages and lead to lower final yiel d (Bhattacharya, 2019; Petersen, 2019; Van Oort and Zwart, 2017)

PAGE 29

29 Food crops, including rice, are predicted to be influenced by climate change across many regions of the world with different impacts among the climate factors. The increased CO 2 is expecte d to influence biomass and rice yield positively through photosynthesis effects (Ainsworth, 2008; Usui et al., 2016) and affect the quality of the grain negatively regarding nutrition content (Myers et al., 2014) The impacts of temperature and CO 2 on rice growth have been investigated in controlled experiments and models have been created to perform evaluations of different climate scenarios (Bhattacharya, 2019; Chaturvedi et al., 2017; Haile et al., 2017; Qiao et al., 2019) Both experimental evaluati ons of climate change under controlled conditions and work using crop models indicate that the net rice yield is expected to decrease under a combined increase in temperature and an increase in CO 2 concentration in the atmosphere (Singh et al., 2014; Wang et al., 2015, Wang et al., 2016) Rice is grown in all ten provinces of Haiti. However, over 70% of Haitian rice is produced in the Artibonite Valley, 15% of the rice production is found in the Northern region, while the rest is cultivated in the Southwest of the country. Total rice acreage in the Artibonite Valle y is over 50,000 ha with some of these areas only cultivated during the rainy season due to lack of available water for irrigation during the dry season (Wilcock and Jean Pierre, 2012) The crop is cultivated mostly under flooded condition s or under irri gation during all three seasons (spring summer, summer autumn and winter spring). About six varieties of rice are grown in the Valley, including Malaka, Schela Bogapot, Tididi, CAP and TCS10. The cultivars TCS10, CAP and Malaka are produced for their y ield while the variety Schela is grown for its culinary quality. Rice is planted in April for the spring summer, August for the summer autumn and December for the winter spring season. The spring summer season accounts for 60 70 percent of

PAGE 30

30 total production Ressources Naturelles et du Developpement Rural, 2015) their income and livelihood. Due t crop production has been seriously impacted, causing a reduction in crop yield and an increase in food insecurity ( Cohen and Singh, 2014) Current agricultural management practices, such as crop varieties and planting dates, are challenged by the unpredictable rainfall and the already weak agricultural infrastructure impacted by extreme weather events. S ince 1960 t he average annual precipitation has decreased by 5 mm per month per decade and the frequency of hot days and nights has increased from 48 to 63 days between 1960 and 2003 (USAID, 2013) Climate change will likely exacerbate these conditions Many researchers have investigated the potential effects of the chang ing cl imate on agricultural production on global and regional scale s (Challinor et al., 2014; Dixit et al., 2018; Petersen, 2019; Reilly et al., 2003; Tubiello et al., 2002; Xiong et al., 2008) They found that agricultural crops would be affected in some regi ons while in other locations the changing climate might be beneficial to crop production. However, most of these studies do not generally consider seasonal yield variation with the climate. Only a few studies (ME, 2013; USAID, 2011; USAID, 2013; USAID, 20 17) have focused on predicting the potential impact of climate change in Haiti and indicate d that climate factors we re expected to exacerbate the living condition hardships and have a severe impact on the Haitian agriculture. However, these studies were conducted on a national scale with no regard for regional or local assessment and did not evaluate the impacts of the changing climate on crop yield Also, these simulations did not consider the local soil nor the commonly grown crop cultivars. The potential impact s of climate

PAGE 31

31 change on rice yield remains therefore unclear in the region of the Artibonite Valley of Haiti. Based on earlier climate change studies (Pranuthi and Tripathi, 2018; Tiepolo & Bacci, 2017; USAID, 2011) and the last Intergove rnmental Panel for Climate Change (2014) this study hypothesizes that climate change will reduce rice yield for the Artibonite Valley in Haiti. The objectives of this study were to simulate, on a seasonal basis, the future climate trend for the near ter m (2010 2049) and the midcentury (2050 2069) climate periods under both the RCPs 4.5 and 8.5, to evaluate the impacts of these changes on rice productivity and to determine the main climate factors influencing the decrease of rice yield in the Artibonite V alley of Haiti. To investigate these objectives, the Crop Environment Resources Synthesis (CERES) R ice model of the Decision Support System for Agrotechnology Transfer (DSSAT) cropping system was used. The model was calibrated and evaluated with rice exper iment data from the Artibonite Valley and the impacts of the changing climate on rice yield were simulated with five climate models and current management practices of the rice farmers in the Valley. Materials and M ethods Study Site The study was conducted in the Artibonite Valley located between 1850'' and 1918'' N latitude and between 7237'' and 7200'' W longitude in Haiti (Figure 2 1). Rice is cultivated in over 35,000 ha in the Valley under flooded conditions for all growing s easons. About 18,000 ha are cultivated during the rainy season, due to unavailability of water in these locations in the dry seasons. Approximately 60 km of stream cross the regions and the water distribution and the irrigation infrastructures are operated by the Organization for the Development of the Artibonite Valley (ODVA), one of the largest operating bodies of the Ministry of Agriculture (Wilcock and Jean Pierre, 2012)

PAGE 32

32 The maximum temperatures ranged from 29.0C in December to 34.3C in July and the minimum temperatures varied between 17.8 C in January and 19.9 C in July. The average annual temperature was 27C and the average relative humidity was 63% in April and 69% in September. There is a rainy season from May to October which receives 500 to 1200 mm of rain and a dry season from November to April which receives 50 to 100 mm of rain (Lamy et al., 2011) However, the study did not provide the period of record from whi ch these averages were obtained. The NASA Prediction of Worldwide Energy Resources (POWER) provides weather data starting from 1981. We used NASA POWER to estimate the average temperature of the region. The average temperature in the Artibonite Valley was estimated from 1981 to 2017. The average maximum temperatures in the location were 29.6 C in December and 34.0 C in July and the minimum temperatures varied from 19 C in December to 23.8 C in August. The results are similar to the previous literature. Weather D ata for Model Input In Haiti, temporal weather station data are limited and were not available for th is study area. Satellite data was an alternative weather data source although they have inherent uncertainties. For locations with limited meas ured data, gridded weather data can be used to meet data needs for crop simulation models (Battisti et al., 2018) Bai et al (2010) used NASA POWER satellite solar radiation data and ground station data to simulate potential corn yield in China. They concluded that NASA solar radiation data is a viable option for the regional and national crop simulation model. The NASA POWER satelli te weather system ( https://power.larc.nasa.gov ) was developed to make meteorological data available for direct use in fields such as architecture, energy and agrometeorology. Weather information is derived from g ridded data systems and multiple data

PAGE 33

33 sources directly (Maldonado et al., 2019) The weather data of the last version of NASA POWER (POWER release 8) are available on a global grid with a spatial resolution of 0.5 o latitude by 0.5 o longitude (Stackhouse et al., 2018) Kontgis et al. (2019) used NASA POWER to calibrate the CERES Rice in the Mekong River Delta in Vietnam as there were no onsite weather data available. Therefore the NASA POWER weather daily rainfall, temperature (max and min) and solar radiation were used for the calibration of the model applied in this study for the Artibonite Valley of Haiti. Experimentally C ollected D ata for the C rop Three years of crop data including management practices, phenolog y and yield for two varieties (TCS10 and CAP) were provided by the Ministry of Agriculture of Haiti. These experiments were conducted to identify a fertilizer recommendation for rice in the Artibonite Valley for the farmers. The researchers were looking at the rice grain yield response of di fferent varieties including CAP and TCS10 to different nitrogen application levels (0, 30, 50, 70, 75, 90, 100, 120 and 150 kg N ha 1 ). The pre germinated seeds were sown in shallow furrows ; then seeds were covered with a thin soil layer. The length of stay of seedlings in the nursery was 22 to 25 days. Before the trial was set up, the soil was plowed and harrowed with a tiller. The two variety experiments provided by the Ministry of Agricult ure of Haiti were conducted in randomized block design at an experimental farm (Mauger) in the Artibonite Valley. The first experiment was conducted in the winter summer season of year 2012 and had three blocks and four treatments (0 kg N ha 1 50 kg N ha 1 75 kg N ha 1 100 kg N ha 1 ) Each experimental unit was subdivided into two subunits in which the two rice cultivars (TCS10 and CAP) were assigned. The second experiment (2014) had four blocks and six treatments ( 25 kg N ha 1 50 kg N ha 1 75 kg N ha 1 100 kg N ha 1 125 kg N ha 1 and 150 kg N ha 1 ) and was conducted in summer autumn with only the TCS10 cultivar The third experiment conducted

PAGE 34

34 with the TCS10 variety in spring autumn (2015) had three blocks with four treatments (Table 2 1) in each block. Two experiments were conducted with the CAP cultivar alone The first experiment was conducted in autumn winter of the year 2012 in three blocks with four treatments ( 3 0 kg N ha 1 60 kg N ha 1 90 kg N ha 1 120 kg N ha 1 ). The second one was conducted in spring autumn season (2016) in three blocks with seven treatments per block (Table 2 2) Soil D ata Different sampling points of a soil survey conducted in the Artibonite Valley have shown that the valley is relatively homogeneous and generally has sandy loam soils with an average of 29% sand, 30% silt and 41% clay between 0 and 30 cm deep; 29% sand, 31% silt and 40% clay between 30 and 60 cm deep. These soils are neutral to slightly alkaline with pH between 7.3 and 8.0 in layer 0 and 30 cm and 7.3 and 8.0 in layer 30 at 60 cm depth. This alkalinity is due to an accumulation of calcareous alluvium forming the soils of this valley. The average organic matter 1.7% in layer 0 and 30 cm and 1.4% in layer 30 and 60 cm deep. The electrical conductivity is low, i.e. 0.54 mmhos/cm between 0 and 30 cm and 0.60 mmhos/cm between 30 and 60 cm deep. At the upper soil horizon (0 30 cm), the nitrogen content is on average 6.75 t/ha; the assimilable phosphorus content is averaged to 0.050 t/ha; and the potassium is 0.42 t/ha. The results between 30 and 60 cm deep showed that the average nitrogen content is 4.172 t/ha; the average assimilable phosphorus is 0.058 t/ha; and the potash average content is equal to 0.421 t/ha (Louissaint and Duvivier, 2003) Lamy et al. (2011) performed soil analysis at the experimental farm where the rice experiments were conducted and repo rted data for two layers (0 30 cm; 30 60 cm) (Table 2 3).

PAGE 35

35 DSSAT CERES Rice The DSSAT cropping system model is an advanced process oriented and well developed physiologically based crop growing dynamic simulation over time and space. The DSSAT (version 4.7 ) CERES Rice model was used in this study. The optimal range of temperature considered in the CERES effect outside of this range of temperature (Ritchie et al., 1998) The model computes daily phenological development and the biomass production in response to the components of the soil climate and crop management practices (Tian et al., 2018) The input data files required by DSSAT are weather (temperature minimum, temperature maximum, rainfall, solar radiation), soil ( soil type, organic matter, texture, electrical conductivity etc.) genotype characteristics (cultivar) and management practices (crop variety, planting and transplanting date and harvesting date, sowing depth, nitrogen f ertilization and irrigation, row spacing, etc.). Calibration and Evaluation of the M odel Data used to calibrate and evaluate the model are from experiments conducted by the researchers of the Ministry of Agriculture of Haiti. The field experiments were completed in flooded conditions with no water stress and different levels of nitrogen were applied in each experiment. Model ca libration for CSM CERES Rice application in Haiti was performed by fitting genetic coefficients to reduce variation between observed and predicted values for phenology, biomass and yield The genetic coefficients contain the phenological (P1, P2O, P2R, P5) and the growth (G1, G2, G3, G4) coefficients (Table 2 4). Two years of data (2012 and 2014: TCS10 cultivar; 2012 (winter summer season) and 2012 (autumn winter season): CAP cultivar) of rice yield, phenology and management practices were used to estimate the genetic parameters. The second experiment with the CAP cultivar ended in winter 2013. One year (2015: TCS10 cultivar;

PAGE 36

36 2016: CAP cultivar) of data was used for model evaluation The estimation of the phenological coefficients was performed with the ant hesis, and the growth coefficients were estimated with the physical maturity and the final yield. The Generalized Likelihood Uncertainty Estimation (GLUE) (Beven and Binley, 1992) is a Bayesian Monte Carlo technique which has been widely used for model p arameter estimation (He et al. 2010; Rankinen et al. 2006) and is suitable for highly parameterized models (Sun et al., 2016) The DSSAT cropping system has incorporated the GLUE tool for estimating the cultivar coefficients for growth and phenology with several treatments and environments (Jones et al., 2011) The model calibration for the TCS10 cultivar was completed with two treatments (100 kg N ha 1 and 150 kg N ha 1 ) with the highest nitrogen application and without water stress from two different experiments (2012 and 2014, respectively). The first experiment was conducted in winter spring 2012 and the second in summer winter 2014. The model evaluation was then perfor med using 120 kg N ha 1 nitrogen treatment with no water stress of an experiment conducted in summer autumn of the year 2015. The calibration of the model for the CAP cultivar was completed with both 120 kg N ha 1 nitrogen treatment conducted in the winte r spring season and 100 kg N ha 1 nitrogen of autumn winter season of the year 2012. These two treatments were free of water stress and were the highest nitrogen amount applied. The evaluation of the model was done using a 150 kg N ha 1 treatment with no w ater stress from the spring autumn season of 2016. CSM CERES Rice performance consisted of estimating the genetic coefficients of the cultivars from the experimental crop data and assessing the difference between the simulated and the observed growth and p henolog y The GLUE tool was simulated 30,000 times to estimate the

PAGE 37

37 genetic parameters in the calibration process. The generated genetic coefficients were then used to simulate the model and assess the difference between the observed and the predicted pheno log y and growth values. When the results showed a suitable agreement between the predicted and the observed values, the evaluation was performed The evaluation was performed using the same genetics parameters to compare the observed and the simulated gr owth and phenolog y values for the treatment (highest nitrogen and no water stress) of the one year experiment The Root Mean Square Error (RMSE) and the Normal Root Mean Square Error (NRMSE) were the goodness of fit indicators used to evaluate model perfor mance by estimating the errors between the observed and the simulated values for rice flowering duration, maturity and yield. 1 1 1 2 S i and O i are the predicted and observed parameter values, respectively; n is the number of observations, is the observed mean value and i is each observation. The simulation is considered excellent when the NRMSE is less than 10 %, good with an NRMSE between 10 and 20 %, fair with an NRMSE between 20 and 30 % and poor if the NRMSE is higher than 30 % (Rinaldi et al. 2003) Climate D ata The Markov weather simulator ( http://gismap.ciat.cgiar.org/MarkSimGCM/ ) was used to generate the climate data using 17 Global Circulation Models in which five GCMs (GFDL

PAGE 38

38 ESM2M, HadGEM2 ES, IPSL CM5A LR, MIROC ESM CHEM and NorESM1 M) data were selected under the RCPs 4.5 and 8.5 ( Table 2 5). These fives climate models are from the phase 5 of the Coupled Model Inter comparison Project (CMIP5) and were the only ones to be bias corrected and downscaled by the Inter Sectoral Impact Model Intercomparison (ISI MIP) (Li et al., 2016; Warszawski et al., 2014) Therefore these GCMs were suitable to be adopted in this study The MarkSim tool was built to produce weather data for Latin America and Africa and the model has been fit to data from about 9200 stations providing daily data around the world. The MarkSim product is based on a third order Markov process that considers events that occur over the previous three days. The range of climate data (solar radiation, minimum and maximum temperatures and rainfall) from 1980 to 2009 downloaded from MarkSim was used as a baseline (Jones and Thornton, 2000) Data from 2010 2039 w ere used for the near term period and data from 2040 2069 was used for the mid century period. The data for the future climate periods were downloaded under the RCP 4.5 and RCP 8.5. The CO 2 concentration levels (baseline [CO 2 ]: 360 ppm; RCP4.5 near term [C O 2 ]: 423 ppm; and RCP4.5 mid century [CO 2 ] : 499 ppm; RCP8.5 near term [CO 2 ]: 432 ppm; and RCP8.5 mid century [CO 2 ] : 571 ppm) were taken from a report of the Agricultural Model Intercomparison and Improvement Project (AgMIP) (Rosenzweig et al., 2018) Model S cenario S imulations In an agricultural system, there are numerous ecophysiological processes influenced by environmental conditions, including CO 2 concentrations, temperature, nutrients, water and management practices. For this study, factors such a s cultivars, soil conditions and management practices remained constant; while temperature, rainfall, solar radiation, and CO 2 concentrations were considered as independent variables.

PAGE 39

39 The MarkSim climate data were treated and organized with the notepad ++ software and imported in the DSSAT weather tool with the extension (.WTG) which means weather generated. The CO 2 concentration values were added in the environmental section of the DSSAT seasonal files. With cultivars, climate data and soil data configured the model simulated the likely impacts of climate change on rice yield under the RCPs 4.5 and 8.5 in the study region. Rice farmers in the Artibonite Valley plant three rice seasons, which are spring summer, summer autumn and winter spring (Table 2 6 ). Most of the farmers transplant the rice the first week of April for the spring summer, mid August during the summer season and the first week of December during the winter spring season (FEWS NET, 2018) Therefore, the planting dates April 15 th August 2 2 th and December 1 st were simulated for the three seasons spring summer, summer autumn and winter spring, respectively. Rice yield and phenolog y (the two cultivars) were simulated for all the seasons during the baseline, the near term and the mid century climate periods under both the RCPs 4.5 and 8. The effects of climate change on the rice growth stages (anthesis and physiological maturity) and y ields were assessed by comparing the outputs of the simulations of the two periods (near term and mid century) under both the RCP 4.5 and 8.5 scenarios to ones obtained during the baseline. The variables of interest (anthesis and maturity duration, and yie ld) were averaged on 30 years for each climate period. Results Calibration and E valuation of the M odel The DSSAT v4.7.1 cropping system model was used to investigate the climate change impacts on rice yield in the Artibonite Valley. For the TCS10 cultivar calibration, the RMSE was 193.4 kg ha 1 for the yield. There was no difference between the observed and the simulated anthesis days (80 days) or the physiological maturity day (108 days). In the evaluation, there was

PAGE 40

40 one day difference in the flowering dates (81 days: observed, 80 days: simulated) and no difference in the maturity dates (106 days). The RMSE for the yield was 144 kg/ha. The calibration results of CAP variety indicated no difference between the observed and the predicted flowering days (78 days) and showed a 1 day difference between the experimental and the simulated maturity days (106 days: o bserved; 107 days: simulated). The RMSE was 505.8 kg ha 1 for the yield. The evaluation indicated no difference between the anthesis (80 days), one day difference in maturity days (107 days: observed; 108 days: simulated) and the RMSE was 210 kg ha 1 for t he rice yield. The DSSAT CERES Rice simulated rice production in the Artibonite region of Haiti for the TCS10 and CAP rice cultivars and showed good agreement to measured data. The NRMSE was 1%, 1% and 2.2% for the anthesis, maturity and yield respectivel y with the TCS10 cultivar and 0%, 1% and 3.5% for the same variables with the CAP cultivar. The NRMSE was under 10% for all the parameters (Jamieson et al., 1991; Rinaldi et al., 2003) Hence, the performance of the model was considered excellent. Table 2 7 shows all the estimated genetic coefficients for the cultivars and Table 2 8 indicates the calibration and the evaluation results. Figur e 2 2 show s the relationship between the simulated and observed rice grain yield for both the TCS10 and CAP cultiva rs in the Artibonite Valley of Haiti. Future C limate S cenarios The future climate scenarios were evaluated considering the changes in the minimum and maximum temperatures the solar radiation and the rainfal l Both temperatures increased with all the GCMs during all seasons for the near term and the mid century climate periods under both the RCPs 4.5 and 8.5 (Figure 2 3 ) compared to the baseline. Under the RCP 4.5, the increase of the ensemble mean maximum tem perature during the spring summer season was 1.2 1.6 o C and 1.7 2.3 o C for the near term and the mid century climate periods, respectively. Using the same

PAGE 41

41 RCP and for these same climate periods correspondingly, the ensemble mean maximum daily temperature i ncreased by 1.6 2.0 o C and 2.0 2.8 o C during the summer autumn and by 0 1 0.5 o C and 0.7 1.3 o C during the winter spring Under the RCP 8.5, the ensemble mean of maximum daily temperature increased during the three seasons (spring summer, summer autumn and winter spring respectively) by 1.3 1.6, 1.7 2.1 and 0.2 0.6 o C for the near term and by 1.9 3.0, 2.4 3.5 and 1.0 2.0 o C during the mid century. Under the RCP4.5 and during the spring summer, the ensemble average minimum daily temperature increased by 1.3 1 .6 o C for the near term and 1.7 2.4 o C for the mid century compared to the baseline. For both the near term and the mid century, respectively, the ensemble mean minimum temperature increased by 1.2 1.9 o C and 1.7 2.8 o C during the summer autumn and rose by 0.1 0.4 o C and 0.4 1.4 o C during the winter spring season. Under the RCP 8.5, the trend was similar for the ensemble mean of the daily minimum temperature that increased in the three seasons (spring summer, summer autumn and winter spring correspondingly) by 1.3 1.7, 1.3 1.9 and 0.2 0.5 o C for the near term and by 2.0 3.0, 2.1 3.4 and 0.8 2.2 o C for the mid century (Figure 2 3 ) The solar radiation with all fives GCMs under the two RCPs (4.5 and 8.5) decreased during the first season (spring summer) and increased during the second (summer autumn) and the third (winter spring) seasons. Under the RCP 4.5, the ensemble mean of solar radiation during the spring summer was predicted to decrease by 0.9 1.0 MJ m 2 day 1 and 0.4 0.6 MJ m 2 day 1 for the near term and the mid century, respectively. During the summer autumn and the winter spring seasons, respectively, the ensemble mean solar radiation was predicted to increase by 1.9 2.4 MJ m 2 day 1 and 0.4 0.7 MJ m 2 day 1 for the near term and by 1.8 2.3 MJ m 2 day 1 and 0.3 0.8 MJ m 2 day 1 for mid century. Under the RCP 8.5, the ensemble average solar

PAGE 42

42 radiation during the spring summer decreased by 0.8 1.1 MJ m 2 day 1 and 0.2 0.7 MJ m 2 day 1 for the near term and the mid century, respective ly. During the summer autumn and the winter spring respectively, the ensemble average solar radiation increased by 2.0 2.3 MJ m 2 day 1 and 0.6 0.8 MJ m 2 day 1 for the near term and by 1.8 2.4 MJ m 2 day 1 and 0.3 0.8 MJ m 2 day 1 for the mid century (Figure 2 4 ). The rainfall was predicted to decrease for two of the three seasons with the five GCMs under both the RCPs 4.5 and 8.5 for the near term and the mid century climate periods. Considering the mean of all the GCMs, the ensem ble mean rainfall is expected to increase during the spring summer season by 30 114 mm and 19 104 mm for the near term and the mid century, respectively. However, for these same climate periods correspondingly, the ensemble average rainfall was predicted t o decrease by 61 89 mm and 38 85 mm during the summer autumn and by 13 39 mm and 12 28 mm during the winter spring under the RCP 4.5. Under the RCP 8.5, the ensemble average rainfall in the spring summer season increased by 60 83 mm and 44 106 mm for the n ear term and the mid century, respectively, compared to the baseline. During the other two seasons (summer autumn and winter spring, respectively), the ensemble average rainfall was predicted to decrease by 63 93 mm and 14 23 mm for the near term and by 35 110 mm and 7 44 mm for the mid century (Figure 2 4 ). Simulation of C limate C hange I mpacts on R ice P roduction The simulation results showed a reduction in the ensemble average of the phenolog y duration for each of the cultivars in this study for the fut ure climate periods compared to the baseline. The rice yield was predicted to increase during the winter spring season while decrease d in spring summer and summer autumn seasons.

PAGE 43

43 Changing climate effects on phenolog y of TCS10 Modeling results predicted t hat the TCS10 rice flowering duration shortened compared to the baseline considering climate change scenarios under the RCP 4.5. During the spring summer, the ensemble mean flowering duration decreased by 4.5 5.5 days for the near term and by 6.1 7.8 days for the mid century. For the near term and the mid century, respectively, the ensemble mean anthesis duration is predicted to shorten by 3.6 4.7 days and 4.5 6.5 days during the summer autumn and by 0.5 1.6 days and 2.3 5.3 days during the winter spring season. Under the RCP 8.5, the ensemble mean of the flowering duration during the spring summer season was predicted to decrease by 4.7 5.7 days for the near term and by 6.8 9.2 days for the mid century. For the two climate periods (near term and mid cent ury, respectively), the ensemble mean flowering duration was predicted to decrease by 3.8 4.8 days and 5.4 7.6 days during the summer autumn season and by 1.0 1.9 days and 3.7 7.9 days during the winter spring season (Figure 2 5 ). The maturity duration also was predicted to shorten during the future climate periods with all the five GCMs. Considering the RCP 4.5 for the near term and the mid century, respectively, the ensemble average physiological maturity decreased by 5.2 6.7 days and 7.4 9.5 days during the spring summer, 6.2 8.2 days and 8.3 10.8 days during the summer autumn and 0.3 3.0 days and 3.7 7.8 days during the winter spring. Under the RCP 8.5 and for the three seasons (spring summer, summer autumn, and winter spring, re spectively) the ensemble average physiological maturity days decreased by 5.4 6.8, 6.8 8.6 and 0.6 3.4 days for the near term and by 8.0 11.4, 9.2 12.4 and 5.8 11.0 days for the mid century (Figure 2 5 ). Changing climate effects on phenology of CAP The CA P cultivar flowering duration was predicted to decrease as compared to the baseline. For the near term and the mid century, respectively, the ensemble average flowering

PAGE 44

44 duration reduced by 4.6 5.7 days and 6.1 8.1 days during the spring summer, by 3.2 4.4 days and 3.4 5.9 days during the summer autumn and by 0.2 1.2 days and 1.8 5.0 days during the winter spring under the RCP 4.5. Under the RCP 8.5, the ensemble mean of the anthesis days during the spring summer decreased by 5.0 5.9 days for the near term a nd by 6.7 9.4 days for the mid century. For the near term and mid century, respectively, the ensemble average anthesis duration shortened 3.4 4.7 days and 5.0 7.1 days during the summer autumn and by 0.9 1.7 days and 3.3 8.0 days during the winter spring ( Figure 2 5 ). The physiological maturity duration also decreased for the future climate periods with all the GCMs used in the study considering the CAP cultivar. For the near term and the mid century, respectively, the ensemble mean physiological maturity decreased by 5.3 6.8 days and 7.7 9.9 days during the spring summer, 6.1 8.1 days and 7.4 8.4 days during the summer autumn and by 0.2 2.7 and 3.4 7.7 days during the winter spring under the RCP 4.5. Under the RCP 8.5 and the mid term and the mid century, respectively, the ensemble average physiological maturity days decreased by 5.8 7.3 days and 8.3 11.6 days during spring summer, by 6.6 8.2 days and 8.9 11.6 days during the summer autumn and by 0.4 3.1 days and 5.3 11.2 days during the winter spring (Figu re 2 5 ). Changing climate effects on rice yield The rice yield for both the cultivars was predicted to decrease during spring summer and summer autumn seasons and increase during the spring winter season for the two climate periods (near term and mid cent ury) under both the RCPs 4.5 (Figure 2 6 ) and 8.5 ( Figure 2 7 ) The ensemble average TCS10 yield was predicted to decrease in two of the three seasons compared to the baseline considering climate scenarios (RCPs 4.5 and 8.5). During the spring summer and t he summer autumn, respectively, the ensemble mean yield was predicted to decrease by 4.9 6. 2 % and 4.4 7.7% for the near term, and by 5. 3 6. 8 % and 4.0 9.1% for the mid

PAGE 45

45 century under the RCP 4.5. Under the same RCP, the ensemble average yield during the wint er spring was predicted to increase by 2.3 4.4% and 1.6 4.1% for the near term and the mid century, respectively. Under the RCP 8.5, the ensemble average yield during the spring summer and summer autumn seasons, respectively, was predicted to decrease by 5 3 6.5% and 5.3 7.6% for the near term and by 5. 8 7.2% and 8.5 11.5% for the mid century. For these two climate periods (near term and mid century, respectively), the ensemble mean of the yield was predicted to increase by 2.0 4.4% and 0.0 4.8 % during the winter spring compared to the baseline. The yield of the CAP cultivar was predicted to decrease in the spring summer and the summer autumn seasons and increase during the winter spring under both the RCPs 4.5 and 8.5. In spring summer and summer autumn se asons, respectively, the ensemble average yield was simulated to decrease by 3.7 4.5% and 3.4 6.4% for the near term, by 4. 7 6.5 % and 3.6 7.3% for the mid century under the RCP 4.5. With the same RCP, the ensemble average yield during the winter spring inc reased by 1.8 3.2% and 2.2 3.2% for the near term and the mid century, respectively. Under the RCP 8.5, the ensemble average of the yield during the spring summer and the summer autumn decreased, respectively by 4.4 5.5% and 4.0 6.1% for the near term and by 4. 6 5.6% and 4.7 11.2% for the mid century. For these two climate periods (near term and mid century respectively), the ensemble average yield increased by 1.7 2.9 % and 0.4 3.8% during the winter spring. Discussion Model P erformance The quality of the data used in this study to calibrate and evaluate the model was considered as intermediate according to the normal quality standards (Kersebaum et al., 2015) Information such as initial soil nitrogen condition was missing in the datas et; hence, assumptions were made to estimate the initial soil nitrogen concentrations based on soil data for the site and

PAGE 46

46 experimental rice yield data. Information related to management practices such as transplanting date, application date and rate of fer tilizers, anthesis and maturity duration and final yield were sufficient to perform proper calibration and evaluation of the model according to the required minimum data set (Bao et al., 2017; Dos Santos et al., 2016). The rice phenologies (flowering da te and maturity date) and yield were well calibrated and evaluated. The available field data for calibration and evaluation was relatively small as only two treatments from two different experiments were used for the calibration and one treatment of anoth er experiment was used to evaluate the model. The Normalized Root Mean Square Error (NRMSE) of the phenolog y and the yield of both the cultivars was less than 10% indicating a good performance with the model. A similar level of NRMSE lower than 10% was fo und in previous studies using CSM CERES Rice for rice flowering duration, physiological maturity duration and grain yield with CERES Rice (Basso et al., 2016; Li et al., 2016; Zhang et al., 2019) Model S imulation The number of days required for floweri ng and for physiological maturity to occur was predicted to decrease with the two cultivars for the future climate periods for all the seasons due to the increase of temperature. However, the anthesis and the maturity day reduction was greater with the CAP cultivar as compared to the TCS10 cultivar. Considering an average of all three seasons, the ensemble average of the flowering and maturity duration for the TCS10 cultivar was not significantly different from the CAP variety. Although there were no previo us studies of climate change with these rice cultivars in the Artibonite Valley in Haiti, similar findings in other locations were reporte d with models predicting a reduction in anthesis and maturity duration for rice due to rising temperature. For instanc e, Shi et al. (2015) found a decrease of the phenological stages due to heat stress using RiceGrow model; Xu et al. (2017) came out with

PAGE 47

47 similar conclusions assessing climate change impacts on rice with CSM CERES Rice in the Sichuan basin in China Dur ing the winter spring season, the increase of the temperature triggered the increase of the yield compared to the other seasons. The average temperature during the winter spring season did not exceed the optimal range of temperature (14 o C to 32 o C) favorab le to yield in the CERES model. Hence, the rice yield was expected to increase during that season although the yield usually recorded is the lowest one among the seasons. Rice yield was predicted to benefit from the changing climate in winter spring even w ithout the inclusion of the CO 2 fertilization effects. Rice is considered a heat stress tolerant crop and has a higher optimal temperature for growth compared to temperate and sub tropical cereal crops. In the tropics, the increase in the local temperatures is usually smaller than the global mean temperatures (Iizumi et al., 2017) Hence, the relatively low temperature during the winter spring season for the future climate periods might benefit r ice production in the Artibonite Valley. The rice yield (for both the cultivars) during the spring summer and the summer autumn seasons was predicted to decrease due to the warming temperature in the Artibonite Valley. Ritchie et al. (1998) stated that t he reduction of the average growth rates and growth durations due to the rising temperature outside the range could significantly affect rice yield. The net photosynthesis, which is the difference between photosynthesis and respiration, govern the growth a nd is influenced by the temperature whereas warm temperature would increase the biomass consumption by increasing the respiration (Hatfield et al., 2011) Also, the warming climate could trigger lower yield, spikelet sterility and a potential crop failur e when close to the extremely high temperature, especially during the anthesis growth phase (Jgadish et al. 2015; Lobell and Gourdji, 2012) Thus, the increase of the temperature close to very high temperatures

PAGE 48

48 during the rice growing seasons (spring su mmer and the summer autumn in the Artibonite Valley) would reduce the rice growth rate by reducing the net photosynthesis. The growth duration, which includes flowering and maturity duration that determines the time for grain filling and biomass accumulati on and was predicted to decrease under the RCPs (4.5 and 8.5) of the changing climate. Therefore, rice yield for both the cultivars was expected to decrease. This finding was consistent with previous climate studies such as Bocchiola et al. (2015) Krish nan et al. (2007) and Xu et al. (2017) who predicted a reduction of rice yield using PolyCrop, ORYZA1 and I nfo C rop and CERES Rice models, respectively. Conclusion This study illustrated the potential impacts of climate change on rice yield in the Arti bonite Valley, which is the main area of rice production in the country. The impacts of the future climate for the near term and the mid century under the RCP 4.5 and 8.5 were investigated using the CSM CERES Rice model. Rice yield was predicted to decreas e during two of the three seasons (spring summer and summer autumn) under both the RCPs 4.5 and 8.5. However, the ensemble average yield, during the winter spring season, was predicted to increase under the RCP 4.5 and 8.5, compared to the baseline. The re sults suggested that adaptation strategies should be assessed to address the potential negatives effects of the changing climate on rice yield during the spring summer and the summer autumn seasons and optimize the rice production in the Artibonite Valley.

PAGE 49

49 Figure 2 1 Region of rice production in the Artibonite Valley (MARNDR, 2013)

PAGE 50

50 Table 2 1. Input parameters of management practices of the TCS10 cultivar for the DSSAT CERES Rice model. These data are from field experiments conducted in the Artibonite Valley of Haiti by the Ministry of Agriculture of Haiti. Variety: TCS10 (2012) Application dat e and proportion of fertilize r (urea) applied Treatments Transplanting date and season 20 days AT 31 days AT 62 days AT Harvesting date and seasonS 0 kg N ha 1 02/16/2012 winter 0% 0% 0% 06/03/2012 summer 50 kg N ha 1 30% 40% 30% 75 kg N ha 1 30% 40% 30% 100 kg N ha 1 30% 40% 30% Variety: TCS10 (2014) Application date and proportion of fertilizer (urea) applied Treatments 15 days AT 30 days AT 61 days AT 25 kg N ha 1 07/30/2014 summer 30% 40% 30% 11/29/2014 autumn 50 kg N ha 1 30% 40% 30% 75 kg N ha 1 30% 40% 30% 100 kg N ha 1 30% 40% 30% 125 kg Nha 1 30% 40% 30% 150 kg Nha 1 30% 40% 30% Variety: TCS10 (2015) Application date and proportion of fertilizer (urea) applied Treatments 30 days AT 45 days AT 58 days AT 30 kg N ha 1 05/13/2015 spring 30% 40% 30% 09/17/2015 autumn 60 kg N ha 1 30% 40% 30% 90 kg N ha 1 30% 40% 30% 120 kg N ha 1 30% 40% 30% AT: After transp l antation

PAGE 51

51 Table 2 2. Input factors of management practices of the CAP variety for the DSSAT CERES Rice model. These data are from field experiments conducted in the Artibonite Valley of Haiti by the Ministry of Agriculture of Haiti Variety: CAP (2012) Application date and pr oportion of fertilizer (urea) applied Treatments Transplanting date and season 20 days AT 31 days AT 62 days AT Harvesting date and season 0 kg N ha 1 02/16/2012 winter 0% 0% 0% 06/03/2012 summer 50 kg N ha 1 30% 40% 30% 75 kg N ha 1 30% 40% 30% 100 kg N ha 1 30% 40% 30% Variety: CAP (2012) Application date and proportion of fertilize r (urea) applied Treatments 28 days AT 42 days AT 56 days AT 30 kg N ha 1 09/24/2012 autumn 30% 40% 30% 01/03/2013 winter 60 kg N ha 1 30% 40% 30% 90 kg N ha 1 30% 40% 30% 120 kg N ha 1 30% 40% 30% Variety: CAP (2016) Application date and proportion of fertilizer (urea) applied Treatments 28 days AT 42 days AT 56 days AT 0 kg N ha 1 05/14/2016 spring 0% 0% 0% 09/30/2016 autumn 25 kg N ha 1 30% 40% 30% 25 kg N ha 1 30% 30% 30% 75 kg N ha 1 30% 40% 30% 100 kg N ha 1 30% 40% 30% 125 kg N ha 1 30% 40% 30% 150 kg N ha 1 30% 40% 30%

PAGE 52

52 Table 2 3. Soil data of the Mauge farm in the Artibonite Valley (Haiti) Profile pH (H2O) EC (Mmhos/cm) K (cmol/kg) C.E.C (cmol/kg) N (%) P (p.p.m) OC (%) OM (%) Clay (%) Silt (%) 0 30 cm 8.2 0.364 0.27 23 0.19 0.14 1.71 2.94 41 30.2 30 60 cm 8.28 0.322 0.32 22.2 0.14 11.5 1.06 1.82 39.8 31.1 EC: Electrical conductivity K: Potassium CEC: Cation exchange capacity N: Nitrogen P: Phosphorus OC: Organic Carbon OM: Organic Matter Table 2 4 Genetic coefficients for rice cultivar and their definitions in the DSSAT CERES rice model (Hoogenboom et al., 2017). Coefficients Definitions P1 Period expressed as growing degree days (C days) above a base temperature of 9 C in the basic vegetative phase of the plant P20 The critical photoperiod or the longest day length (in hours), during which development occurred at a maximum rate P2R The Extent to which phasic development leading to panicle initiation is delayed (expressed as GDD in C) for each hour increase photoperiod above P2O P5 The time period (C days) from the beginning of grain filling to physiological maturity with a base temperature of 9 C in grain filling phase G1 The potential spikelet numbers coefficient per panicle G2 The single grain weight under ideal growing conditions (nonlimiti ng light, water, nutrients and absence of pests and diseases) G3 The tillering coefficients relative to IR64 cultivars G4 The temperature tolerance coefficient

PAGE 53

53 Table 2 5 Information related to the five GCMs used in this research study Table 2 6 Current planting and harvesting dates of the farmers of the Artibonite Valley. These planting dates were simulated in the DSSAT seasonal tool to evaluate the impacts of the climate change on rice yield for each growing rice seasons Seasons Jan Feb March April May June Jul August Sept Oct Nov Dec spring summer P H summer autumn P H Winter spring H P P: Planting period; H: harvesting period Table 2 7. Genetic coefficient values (unitless) for TCS10 and CAP rice cultivar after calibration with GLUE. These two varieties are among the widely cultivated rice cultivars in the Artibonite Valley of Haiti. Genetics coefficients Cultivars TCS10 CAP P1 475.1 638.9 P2R 175.3 195.2 P5 375.5 373.6 P20 11.1 12.2 G1 46.3 48.4 G2 0.025 0.025 G3 0.90 0.82 G4 1.00 0.93 GCM names Model source GFDL ESM2M National Oceanic and Atmospheric Association Geophysical Fluid Dynamics Laboratory, USA HadGEM2 ES Met Office Hadley Center, UK IPSL CM5A LR Simon Laplace, France MIROC ESM CHEM Japan Agency for Marine Earth Science and Technology, Atmosphere and Ocean Research of the University of Tokyo and National Institute for Environmental Studies NorESM1 M Norwegian Climate Centre

PAGE 54

54 Table 2 8. Calibration and evaluation of the cultivars (TCS10 and CAP) in Artibonite Valley in Haiti Calibration TCS10 CAP Variables Obs. Sim. RMSE RRMSE Obs. Sim. RMSE RRMSE Anthesis (days) 80 80 0.7 0.01 78 78 5.523 0.07 Maturity (days) 108 108 1.0 0.01 108 107 4.123 0.04 Yield (kg/ha) 5520 5712 193.4 0.04 5768 6114 505.8 0.09 Evaluation TCS10 CAP Variables Obs Sim RMSE RRMSE Obs Sim RMSE RRMSE Anthesis (days) 81 80 1 0.01 80 80 0 0.00 Maturity (days) 106 106 0 0.00 107 108 1 0.01 Yield (kg/ha) 6375 6519 144 0.02 7974 8184 210 0.03

PAGE 55

55 Figure 2 2. Relationship of observed grain yield with simulated yield of rice for (a) TCS10 and ( b ) CAP cultivars used in the study. These observed data were collected in experimental trial s conducted in the Artibonite Valley of Haiti in 2012, 2 014, 2015 and 2016 by the Ministry of Agriculture of Haiti These data were not used in the calibration process The dashed line represents the line 1:1 and the red line is the regression line.

PAGE 56

56 Figure 2 3 Predicted variation of the maximum and minimum temperatures with five GCMs during the three seasons (spring summer, summer autumn and winter spring) for two future climate periods (near term:2010 2039; and mid century: 2040 2069) under both RCPs 4.5 and 8.5 compared with the baseline. Each of the periods is 30 years of climate data. The fives GCMs are on the top x axis and the three seasons are on the right y axis The green, red and dark red colors indicate the baseline, near term and the mid century climate periods respectively.

PAGE 57

57 Figure 2 4 Predicted variation of the solar radiation and rainfall with GCMs in the three seasons (spring summer, summer autumn and winter spring) for two future climate periods (near term:2010 2039; and mid century: 2040 2069) under both RCPs 4.5 and 8.5 compared w ith the baseline. Each of the periods is 30 years of climate data. The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively.

PAGE 58

58 Figure 2 5 Predicted changes in the flowering duration (a) and maturity duration (b) for two future climate periods (near term and mid century) under two RCPs (4.5 and 8.5) during three seasons (spring summer, summer autumn and winter spring). The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively. The flowering and maturity days for each of the three rice growing seasons are on the left y axis, the cultivars are on the bottom x axis and the GCMs are on the top y axis.

PAGE 59

59 Figure 2 6 Simulated yield ( kg/ha) of the TCS10 and CAP rice cultivars for two climate periods (near term and mid century) during three seasons (Spring summer, summer autumn and winter spring) with the five GCMs in the Artibonite Valley under the RCP 4.5 compared to the baseline. The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively The cultivars (TCS10 and CAP) are in the bottom x axis, the GCMs are on the top x axis and the yield for each of the tree rice growing seaso ns are on the left y axis

PAGE 60

60 Figure 2 7 Simulated yield (kg/ha) of the TCS10 and CAP rice cultivars for two climate periods (near term and mid century) during three seasons (Spring summer, summer autumn and winter spring) with the five GCMs in the Artibonite Valley under the RCP 8 .5 compared to the baseline. The green, red and dark red colors indicate the baseline, near term and the mid century climate periods, respectively The cultivars (TCS10 and CAP) are in the bottom x axis, the GCMs are on the top x axis and the yield for each of the tree rice growing seasons are on the left y axis

PAGE 61

61 CHAPTER 3 SIMULATING POTENTIAL ADAPTATION STRATEGIES TO ALLEVIATE CLIMATE CHANGE IMPACTS ON RICE YIELD IN THE ARTIBONITE VALLEY OF HAITI Introductio n The future Greenhouse Gas (GHG) emissions are projected to continue rising and scenarios show that global mean temperature could increase by 2.6 o C, sea level is expected to rise from 0.22 m to 0.38 m and rainfall is projected to be highly variable across the world by the middle of the 21st century (IPCC, 2014) Haiti is not exempt to these changes and has already experienced rising temperatures of 0.12 o C per decade from 1970 to 2013 (Borde et al., 2015) ; the temperature increase is expected to reach up to 2.3 o C by 2060 (USAID, 2013) The rising temperature impacts agricultural production around the world. Increased temperature can disturb the physiological processes, including respiration and photosy nthesis and can reduce the growth stages and grain filling rate (Boote, 2011; Yang and Zhang, 2006) Nevertheless, a warmer climate could be beneficial in a region where the growing season length is limited by temperature (Shrestha et al., 2016). The poo rer countries are most vulnerable to the consequences of the changing climate due to their weak response capacity. Haiti is the most vulnerable country to climate change in the Caribbean region and the most impoverished country in the American continent wi th a GDP per capita of US $729 in 2019. The GDP is at 8.41 billion dollars and the growth rate is estimated at 1.50% (Trading Economics, 2019). Borde et al. (2015) stated that about 59% of Haitians live under the poverty line with $2.44 per day and 24% l ive below the threshold of extreme poverty set at $1.24 per day. They also highlighted, among other things, the increased water requirements for crops as a result of the combination of rising temperature and rainfall variability. The simulation of these ch anges showed that water needs for corn would increase by 17% and 33%, respectively, for 2011 2040 and 2041 2070, and water needs of rice will increase of 0.2% and 7% over the same

PAGE 62

62 response, these changes are predicted to exacerbate its existing vulnerabilities. Adaptation strategies are crucial to address and reduce the impacts of the chang ing climate. Adaptation measures in agriculture, such as changing planting dates, use of tem perature and water stress resistant cultivars, irrigation management and relocation of agricultural production areas have been proposed to offset the negatives impacts of climate change (Reilly et al., 2003; Rosenzweig et al., 2004; Tubiello et al. 2002) Rosenzweig and Tubiello (2007) reviewed issues concerning climate change and agriculture with an emphasis on adaptation and mitigation. They found that adaptation measures that are needed for maintaining the production and income might not go along wi th the usual local agricultural practices. However, their findings showed that these strategies are essential to reduce the negatives effects of climate change on agriculture. Appropriate climate adaptation actions of crops including rice to climate change have been widely assessed and validated, but they are usually specific to a region based on the many different properties and climate (Bryan et al., 2009) Boonwichai et al. (2019) evaluated the effectiveness of four climate change adaptations strateg ies, viz., shifting planting and fertilizer application date, fertilizer application dose adjustment and increased irrigation water in Thailand. Their results showed that changing sowing date, the application date and fertilizer application amount do not a lways trigger an increase in the yield while changing the irrigation increased rice yield significantly. However, such adaptive actions depend on available resources and the farmer's mindset. Shrestha et al. (2016) found that shifting of transplantation date, irrigation water management, splitting fertilizer doses and the use of new rice cultivars could optimize rice yield under the harmful effects of climate change in central Vietnam. Hence, adaptation

PAGE 63

63 e evaluated regionally and should be assessed on a site specific basis. Climate change impacts on rice cultivation have not been evaluated in the Artibonite Valley, which produces more than 70% of Haitian rice. In chapter 2, the impact of the changing cli mate on rice production was evaluated for all three rice seasons (spring summer, summer autumn and winter spring). The rice yield decreased in spring summer and summer autumn seasons while increased during the winter spring during the near term and the mid century climate periods under both the RCPs 4.5 and 8.5. The spring summer season, followed by the summer autumn, are the most important seasons to the farmers due to the relatively high yield compared to the low harvested yield in the winter spring. The CO 2 increase improved the yield under the changing climate in the seasons but was not enough to remove the total impacts of the rising temperatures in the seasons the yield was predicted to decrease. In this chapter, we hypothesize that implementation of a daptation strategies simulates a reduced impact of the changing climate on rice yield in the Artibonite Valley. However, effective adaptation actions to reduce climate change effects on rice yield in the region are unknown. Therefore, the chapter objective s were to use a crop model to simulate the application of alternative planting dates, different nitrogen application amounts and plant breeding to reduce the impacts of climate change on rice yield in the Artibonite Valley of Haiti. Materials and Methods Experimental S ite The Artibonite Valley is located in the western part of the Republic of Haiti. This Valley River. The rivers cross the whole valley to r each the Gulf of Gonve. The dam of Peligre was built to supply irrigation water to 28,000 ha in the region in addition to other water sources

PAGE 64

64 available. The irrigation system channels flow from the Artibonite River onto valley flood plains. In addition to the Artibonite River, tributaries, springs and lagoons provide water in the region (Wilcock and Jean Pierre, 2012) This set of water resour ces is an asset for growing rice in the valley. There is a rainy season from May to October when the region receives 500 to 1200 mm of rain and a dry season from November to April, with 50 to 100 mm of rain per year (Lamy et al., 2011). The soils are predo minantly calcareous, with a fine to medium texture and are poorly or imperfectly drained. The soil of the experimental site is clay silty with the occurrence of 41% clay. The percentage of organic matter range between 1.49 and 2.94 % and the C/N ratio rema ins low, varying from 3.4 to 9.43 and denotes a low rate of decomposition of organic matter. DSSAT C ropping S ystem M odel Management practices, fertilization, climatic variables and genetics are parameters that influence rice productivity. An optimal combination of these factors would provide the highest possible rice yield. Data collected from long term field experiments tha t were conducted provide information for dynamic simulation models (Liu et al., 2011) Conducting field trials considering all the different variables to measure optimum production combinations is not feasible due to the expense and time required (Vilayv ong et al., 2015) Dynamic crop models are one of the solutions enabling detailed and systematic analyses and fast, low cost and accurate simulation of crop growth and soil nutrient water dynamic (Liu et al., 2011) The DSSAT is a broadly used decisio n support system which contains many application programs which perform seasonal, sequence, crop rotations and spatial simulations. These analyses facilitate the evaluation of environmental impacts and economic risks related to fertilizer and nutrient mana gement, irrigation, soil carbon sequestration, precision management, climate variability and climate change (Hoogenboom et al., 2017) The DSSAT encompasses

PAGE 65

65 crop simulation models for over 42 crops, including Crop Estimation Resource and Environment Synt hesis (CERES) rice (crop model for rice) and tools that enable creation and management of soil and weather data and experimental crop files. The CERES rice model, linked to the DSSAT, is one of the most successful tools which has been used for simulating rice yield and growth under irrigated and non irrigated conditions (Ahmad et al., 2013) Also, DSSAT simulates the rice genetics coefficients, weather, agricultural practices and soil dynamic on the rice development, growth and yield (Timsina and Humphre ys, 2006) and assesses the impact of climate change on rice yield (Yao et al., 2007) Weather, E xperimental R ice and S oil D ata and M odel P erformance Two rice cultivars (TCS10, CAP) widely cultivated in the Artibonite valley were considered to evaluate ad aption strategies. The two rice varieties were calibrated and validated in the DSSAT CSM (chapter 2). Weather data from NASA Prediction of Worldwide Energy Resources (POWER), existing experiment rice data and existing soil data were used to estimate the ge netic coefficients for each of the cultivars. Since there was no weather data available on site where the experiments were conducted, weather data such as temperatures (maximum and minimum), rainfall and solar radiation from the POWER database were used. E xisting soil data (Table 2 3) for two layers (0 30 cm, 30 60 cm) with parameters including pH, electrical conductivity and organic matter from the literature were used in the calibration process of the rice cultivars. Three years of existing experimental rice data were provided with the Haitian Ministry of Agriculture. Two years were used to estimate the genetic coefficients (model calibration) using the Generalized Likelihood Uncertainty Estimation (GLUE) and one year was used for validating the model. T he RMSE and RRMSE were used to evaluate the quality of the calibration and the evaluation of the model (Table 2 8).

PAGE 66

66 Modeling S imulations The stochastic weather generator MarkSim ( http://gismap.ciat.cgiar.org/MarkSimGCM/ ) was used to generate climate under two contrasting greenhouse gas emissions scenarios (RCP 4.5 and RCP 8.5). The Markov weather simulator (Jones and Thornton, 2000) has 17 Global Circulation Models in which five GCMs (HadGEM2 ES, GFDL ESM2M, IPSL CM5A LR, MIROC ESM CHEM and NorESM1 M) were selected under the RCPs 4.5 and 8.5. These scenarios are based on the comparison between the projected radiation forcing by the end of the 21 st century and the pre industrial levels. The RCP 4.5 climate scenario predicts a low medium GHG emission (+4.5 Wm 2 ) and the RCP 8.5 scenario projects a high emission of the GHG (+8.5 Wm 2 ). The climate data (solar radiation, minimum a nd maximum temperatures and rainfall) were grouped in three periods : baseline (1980 to 2009), near term (2010 2039) and mid century (2040 2069). Data of the future climate periods (near term and mid century) were downloaded under the RCP 4.5 and RCP 8.5. T he CO 2 concentration levels considered in the Agricultural Model Intercomparison and Improvement Project (AgMIP) (Rosenzweig et al., 2018) were used for model simulation. Assessment of P otential A gronomic A daptation M easures Multiple adaptation strategi es including shifting planting dates, changing plant spacing, altering irrigation, breeding new varieties (temperature stress tolerant), improving management practices, improving the efficiency of pest and disease control and selection of more pest resista nt cultivars have been evaluated across the globe with crop models (Bhuvaneswari et al, 2014; Chun et al., 2016; Li et al., 2016; Xu et al., 2017a) In this study, the CSM CERES rice was used to simulate agro adaptation measures such as changing transplanting date, altering fertilizers application rate and breeding heat tolerant cultivars.

PAGE 67

67 Changing the transplanting date Changing the transplanting date could improve the crop yield by cultivating the plant during a period of relatively low evaporat ive demand (Mahajan et al., 2009) For evaluating the potential transplanting dates that could reduce the negative influence of rising temperature in the Artibonite Valley, the transplanting dates were advanced and delayed seven times with seven days int erval and were compared to those that are currently adopted in the Artibonite Valley. Fifteen transplanting dates were assessing for each season. However, a transplanting date can be beneficial to yield during a season alone while not necessarily contribut ing to the optimum annual yield with three crops a year. Therefore, transplanting dates during each of the three seasons (which also considered harvest dates) that provided the optimum annual yield was the suitable ones to reach the highest annual rice yie ld. Breeding new rice cultivars For the development of the virtual rice cultivars, the crop life cycle and yield potential traits were considered under the baseline and the climate change scenarios. Three crop maturity durations were identified for develop ing virtual cultivars: origin al cultivar 10% shorter life cycle cultivar and 10% longer life cycle cultivar. Singh et al. (2014) described the procedure for m aking changes in crop maturity duration. To perform changes in the crop maturity duration, we altered the genetic coefficients P1 (thermal time in growing degree days (GDD) from seedling emergence to the end of the juvenile phase), P2O (critical photoperio d or the longest day length in hours at which development occurs at a maximum rate) and P5 (thermal time in GDD from start ing of grain filling to physiological maturity) located in the cultivar file (*.CUL) of the DS S AT CSM. To create the 10% shorter cultivar, the crop duration was decrease d by 10%. P2O was increased and P1 decreased to have 10% reduction in days to anthesis and the value of P5 was

PAGE 68

68 decreased to have an overall 10% reduction in days to physiological maturity duration T he 10% l onger cultivar was created by increas ing P1, decreas ing P2O and increas ing P5 at 10%. To increase the yield potential of cultivars, the coefficients G1, G2 and G3 and the RUE (Radiation Use Efficiency) were increased by 10% each (Singh et al., 2017) The se modifications resulted in six virtual cultivars, namely, or iginal 10% short, 10% long, baseline + yield potential, 10% short + yield potential and 10% long + yield potential However, the drought and heat tolerance genetic traits were not incorporated and were not evaluated Altering Fertilizer Application Rate Fertilizers are a key component in rice production and fertilizer application rate and dose can influence rice yield. Researchers of the Ministry of Agriculture stated that the response of the ri ce variety TCS 10 to nitrogen fertilization in the Artibonite Valley showed that there is no increase in grain yield above 75 kg ha 1 of nitrogen and recommended farmers to not applied more than 75 kg ha 1 However, these recommendations did not reach the farmers, and they continued to apply nitrogen rates of up to 150 kg ha 1 or more (MARNDR, 2014) The fertilizer amount (50, 100, 130, 150, 180, 210, and 210 kg ha 1 ) application were simulated compared to the baseline (75 kg ha 1 ) using the seasonal analysis tool of DSSAT to determine the fertilizer rate that would allow reaching higher rice yield under the two predicted climate scenarios (RCP 4.5 and RCP 8.5) Results Change i n T ransplanting D ates Adjusting the transplanting dates improved the TCS10 cultivar yield during the seasons (spring summer, summer autumn and winter spring) compared to the baseline dates in the Artibonite Valley under the two RCPs (4.5 and 8.5) of the changing climate. Under the RCP 4.5, transplanting dates January 27 th June 3 rd and October 14 th was predicted to increase the annual

PAGE 69

69 TCS10 yield by 12 % for the near term. Likewise, these same transplanting dates ( January 27 th June 3 rd October 14 th ) increase d the annual TCS10 rice yield by 14 % for the mid century under the RCP4.5. U nder the RCP 8.5 the transplanting on January 20 th June 10 th and October 14 th dates predicted an optimum annual yield increase of 1 2 % for the near term and transplanting in January 27 th June 3 rd and October 14 th i ncreased the yield by 15 % for the mid century in the Artibonite Valley ( Figures 3 1 and 3 2 ) The average optimum annual yield was predicted to be reached with the CAP cultivar when transplanting January 27th, June 3rd and O ctober 14th the three rice growing seasons for the near term and the mid century under both the RCPs 4.5 and 8.5. Transplant ing on January 27 th June 3 rd and October 14 th predicted an annual yield increase of 1 2 % for the near term and 13% for the mid century under the RCP 4.5. Under the RCP 8.5, the same transplanting dates ( January 27 th June 3 rd and October 14 th ) predicted a n annual yield increase of 14% in the near term and 1 7 % in the mid century ( Figures 3 1 and 3 2 ) Breeding N ew R ice V arieties The original cultivar TCS10 took 80 days to anthesis and 110 days to physiological maturity and provided an average annual grain yield of 3923 kg ha 1 under the baseline climate with the application of 50 kg N per ha The anthesis and the physiological maturity days o f the 10% short life cycle cultivar w ere 66 days and 94 days, respectively. For the 10% long er life cycle cultivar, the anthesis duration was 97 days and the maturity duration was 128 days. Under the baseline climate, the yield of the 10% short er duration cultivar significantly (p < 0.05) decreased by 20% while the yield of the 10% long longer duration cultivar increased significantly (p < 0.05) by 10 %. When incorporating the 10% increase of yield potential traits under the baseline climate the yield of t he cultivars (original + yield potential, 10% short + yield potential and 10% long + yield potential respectively) increase d by 24%, 25% and 23%

PAGE 70

70 compared to cultivars (original, 10% short and 10% long) that had the original yield p otential traits (Table3 1) Under the RCP 4 5, the annual yield of the 10% shorter maturity cultivar decreased by 19 % for the near term and by 20% for the mid century while the annual yield of the 10% longer maturity cultivar increased by 25 for the near term and by 26% for the mid century, respectively With the 10% increase of the p otential yield traits, the yield increased by 25, 25 and 24 % with the tree virtual cultivars (baseline + yield potential, 10% short + yield potential and 10% long + yield potential, respecti vely) for the near term. For the mid century, the three potential yield traits cultivars increased the yield by 26, 24, and 24%, respectively (Table 3 1 ) Under the RCP 8.5, the average annual yield of the 10% short maturity variety decreased by 20% for t he near term and by 21% for the mid century, respectively while the annual yield of the 10% longer maturity variety increased by 12% in for the near term and by 15% for the mid century. Incorporating the yield potential traits increased the yield by 25, 26 and 24% for the near term and by 26, 24 and 25% for the mid century with the baseline + yield potential, 10% short + yield potential and 10% long + yield potential, respectively cultivars (Table 3 1 ). Altering F ertilizer A pplication R ate The yield of the TCS10 rice cultivar improved with the augmentation of the nitrogen application rate for the near term and the mid century under the both the RCPs 4.5 and 8.5 of the changing climate compared to the b aseline nitrogen application. The nitrogen application rate of 210 kg ha 1 was predicted to trigger an ensemble average yield increase ( 2 2 2 5 % in the near term and 18 23 % in the mid century) during the spring summer season under the RCP 4.5. Under the same RCP, the application of 130 kg ha 1 of nitrogen increased the yield by 1 10% for the near term and 150 kg ha 1 increased the ensemble average by 1 1 4 % for the mid century in the summer autumn T he application of 130 kg ha 1 of nitrogen increased rice yield by 11 26%

PAGE 71

71 and 5 24% for the near term and the mid century, respectively, in the winter spring season. Under the RCP 8.5, the optimum ensemble average yield in the spring summer season was reached with 210 kg ha 1 by increas ing by 2 3 27 % for the near term and by 15 25 % for the mid century The application rate of 130 kg N ha 1 was predicted to increase the summer autumn yield by 3 9% for the near term and 1 5 0 kg ha 1 increased the ensemble average yield by 8 22% for the mid c entury. In the winter spring season, the application rate of 130 kg N ha 1 was predicted to increase the yield by 11 26% for the near term and by 3 30% for the mid century (Figures 3 3 and 3 4 ). Changing of the nitrogen application rate also improved the yield of the CAP cultivar under both the RCPs 4.5 and 8.5 compared to the baseline. The application of 210 kg ha 1 of nitrogen increased the ensemble average yield by 2 2 2 8 % in the near term and 2 7 40 % in the mid century during the spring summer season under the RCP 4.5. With the application of 1 50 kg ha 1 of nitrogen, the ensemble mean yield reached its optimum by increasing by 2 9% for the near term and by 4 14% for the mid century in summer aut umn. The application rate of 130 kg N ha 1 increased the winter spring yield by 11 22% and 5 23% for the near term and the mid century, respectively under the RCP 4 .5. in the winter spring season. Under the RCP 8.5, the optimum ensemble average yield in th e spring summer season was reached with 210 kg ha 1 by increasing by 2 0 2 4 % for the near term and by 18 22 % for the mid century The application of 1 5 0 kg ha 1 was predicted to increase the summer autumn yield by 3 8 % and 1 14 % for the near term and the mi d century, respectively. In the winter spring season, the rate of 130 kg ha 1 was predicted to increase the yield by 11 21 % for the near term and by 3 20 % for the mid century (Figure 3 3 and 3 4 ).

PAGE 72

72 Discussion Shifting transplanting dates to increase yield p redictions with climate change has been identified and widely evaluated in the world (Boonwichai et al., 2019; Kapetanaki and Rosenzweig, 1997; Tingem and Rivington, 2009) The predictions of this study which indicated that early rice transplanting dates for each of the seasons are consistent with many exist ing studies. Xu et al. (2017) evaluated rice planting dates in Sichuan basin in China under the RCP 4.5 and found that advancing planting dates 40 days earlier would increase the rice yield by 24% compared to the current planting but did not consider how these switching dates would affect the other seasons. Kim at al. (2013) performed simulations of rice yield by advancing and delaying the planting dates by 30, 20, and 10 days from the current practices. They concluded that planting 30 days earlier would trigger the highest rice yield under the climate change scenario by the mid century. Babel et al. (2011) evaluated adaptation measures for rice c ultivation in Thailand and concluded transplanting dates 30 days after the current transplanting dates triggered the highest yield by 2020s and 2050s which agreed with the findings of this paper The contribution of the crop life cycle and yield potential were quantified using CSM CERES Rice model under the current and future climate in the Artibonite Valley of Haiti. The changes in temperature and temperatures (min and max) driven by the changing climate will alter the length of the growing periods Achiev ing higher yield with genetic adaptation consists of fitting the rice maturity duration to changes of the length of growing seasons. In that study, the longer yield cycle showed good potential to improve rice yield under the current and the future climates as the warmer climate shorten s the plant life cycle and affects the yield The results are in line with the findings of Singh et al. (2014) who quantified genetic adaptation genetic modification of S orghum as adaptation strategies to climate change using CSM CERES Sorghum Altering yield potential genetic traits (RUE, G1, G2, G3) increased significantly rice

PAGE 73

73 yield. The incorporation of the yield traits improved the yield. However, the baseline + yield potential cultivar provided the highest increased (25% and 26% under the RCP 4.5 and 8.5) compared to the other two cultivars with the yield potential traits. Singh et al. (2017) found similar results with the consideration of the yield potential traits as adaptive measures to the changing climate using CSM CERES Pearl millet and investigated drought and heat tolerance in addition to the evaluation of traits governing crop maturity duration, potential yield According to the findings of this study, a proper cultivar which matches length of growing period is o ne of the best ways to tackle the chang ing climate impacts This could help to minimize the negative impacts of d rought and heat stress during the plant life cycle and the crop would be able to utilize the seasonal resources available abundantly One of the aspects that contrast this paper with the other studies is the consideration of optimum annual yield with the three crop seasons during a year in the Artibonite Valley. Since an optimum planting date that benefited the yield during one season was predi cted to affect the following season, three transplanting dates that trigger the highest annual yield was considered as the optimum transplanting dates strategies under the changing climate in the Artibonite Valley. Changing fertilizers rate improved the yi eld for the future climate periods in the Artibonite valley compared to the baseline. The increase of the amount of the fertilizer (nitrogen) applied augmented the rice yield for the near term and the mid century under both the RCPs 4.5 and 8.5. These conc lusions were in line with the findings of Babel et al. (2011) who concluded that the increase of the fertilizer application rate has the potential to increase the rice yield in the future periods of the changing climate as appropriate fertilizer applicat ion rate is essential for plant growth and yield. Shrestha et al. (2016) simulated climate change adaptations with different fertilizer level in central Vietnam and conclu ded that higher nitrogen input maximized the

PAGE 74

74 potential rice yield. The soil of the Artibonite Valley was predicted to require a relatively high amount of fertilizer to increase the yield under the changing climate. This might be because of the lack of nutrients in the soil. Guo et al. (2017) showed that rice yield increase due to high fertilization application rate would occur in soils where nutrients deficiency is the principal limiting factor. Cochrane et al (2016) stated that the average annual rice yield in Haiti is 2200 kg ha 1 and simulation results of this study indicated tha t the current average annual rice yield is 4874 kg ha 1 when applying the fertilizer rate recommended by the Ministry of Agriculture (75 kg ha 1 ) in the Artibonite Valley. Climate change was predicted to decrease the average annual rice yield by 1 9 71 kg h a 1 by 2069 without adaptations strategies. However, c hanging planting dates was predicted to increase the annual average rice yield by 81 11 1 kg ha 1 by the same and b reeding new cultivars showed the potential to improve the production by increasing the average annual yield by 400 1 096 kg ha 1 compared to the currently cultivated varieties in the Artibonite Valley. Also, the i ncrease of the fertilizer application rate w as predicted to improve the rice yield significantly by increasing the average annual yield by 308 506 kg ha 1 by the year 2069. Conclusion This chapter assessed the effectiveness of the agro adaptations strategies, which included switching transplanting dates, changing fertilizer application rates and plant breeding to remove the climate change effects on rice cultivation in the Artibonite Valley of Haiti. Simulation of transplanting dates allowed to increase the optimum annual rice yield considering thr ee rice seasons a year for both the near term and the mid century under both the RCPs 4.5 and 8.5 of the changing climates. Transplanting planting dates for each of the 42 days for the near term and 28 earlier for the mid century in each of three seasons p redicted the optimum annual yield compared to the baseline. Also, breeding new rice cultivars and changing fertilizer application rate

PAGE 75

75 predicted a significant increase in rice yield for the future climate periods scenarios compared to the baseline. The res ults of these simulations could help the policymakers and other stakeholders to develop and implement adaptation strategies for rice cultivation in the Artibonite Valley. Figure 3 1. Rice transplanting dates for the spring summer, summer autumn and winter spring seasons in the Artibonite Valley under the RCP 4.5 of climate change. The weeks are in the top axis of the graph and the two rice cultivars are in the bottom axis. The green, red and dark red colors represent the baseline, near term and mid c entury climate periods. The dark red boxes are the current planting dates for the spring summer, the summer autumn and the winter spring seasons, respectively.

PAGE 76

76 Figure 3 2. Rice transplanting dates for the spring summer, summer autumn and winter spring s easons in the Artibonite Valley under the RCP 8.5 of climate change. The weeks are in the top axis of the graph and the two rice cultivars are in the bottom axis. The green, red and dark red colors represent the baseline, near term and mid century climate periods. The dark red boxes are the current planting dates for the spring summer, the summer autumn and the winter spring seasons, respectively.

PAGE 77

77 Table 3 1 Grain yield (kg ha ) of virtual rice cultivars derived from TCS10 under baseline climate and projected changes in temperature, CO 2 and rainfall by for the near term and the mid century climate periods at the Artibonite Valley, Haiti. RCP 4.5 RCP8.5 Virtual cultivars Baseline Near term Mid century Near term Mid century Yield %Ch Yield %Ch Yield %Ch Yield %Ch Yield %Ch Original 3922.6 -3874.8 -3884.4 -3881.2 -3916.0 -10% shorter 3153.6 19.6 3125.2 19.3 3122.9 19.6 3125.0 19.5 3083.0 21.3 10% longer 4322.7 10.2 4325.7 11.6 4410.9 13.6 4336.3 11.7 4480.5 14.4 Baseline + yield pot. 4909.3 25.2 a 4853.8 25.3 4877.3 25.6 4866.2 25.4 4921.2 25.7 10% short + yield pot. 3905.6 23.8 a 3897.6 24.7 3868.3 23.9 3892.0 24.5 3823.0 24.0 10% long + yield pot. 5312.8 22.9 a 5378.3 24.3 5487.6 24.4 5390.5 24.3 5576.1 24.5 LSD (0.05) b 0.424 0.181 0.185 0.182 0.188 % Ch: percent yield change of a virtual cultivar due to the trait of crop maturity as compared to the original yield for each climate periods under both the RCP 4.5 and 8.5. a: Percent yield improvement due to enhanced yield potential traits as compared to the cultivar with same crop maturity duration within a climate period. b: Least signi ficant difference (kg ha 1 ) at 5% level of probability to compare yields within the same column or row.

PAGE 78

78 Figure 3 3 Adaptation strategies for change in fertilizer rate application (50, 100,130, 150, 180, 210, and 250 kg N ha 1 ) for the near term (2010 2039) and the mid century (2040 2069) under the RCP 4.5 climate scenario.

PAGE 79

79 Figure 3 4 Adaptation strategies for change in fertilizer dose application (50, 100,130, 150, 180, 210, and 250 kg N ha 1 ) for the near term (2010 2039 ) and the mid century (2040 2069) under the RCP 8.5 climate scenario.

PAGE 80

80 C HAPTER 4 SUMMARY OF FINDINGS AND CONCLUSION Chapter 1 Anthropogenic actions have led to climate change, which has been predicted to affect food security and agriculture around the world. Vulnerable populations relying on agriculture and natural ecosystems in the developing countries have reported being the most affected by the changing climate. T he t emperature increase due to climate change i s predicted to affect crop production and reduce their yield. The Artibonite valley is the main agricultural production area in Haiti and produces more the first and second national climate change rep ort of the country. However, the impacts of climate change on crop yield were not assessed. This study used the DSSAT CERES Rice model and MarkSim GCM to determine the impacts of climate change on rice yield in the Artibonite Valley and provide adaptation measures. The overall goal of this research was to predict the impacts of the changing climate on rice yield in the Artibonite Valley of Haiti and to assess the effectiveness of adaptation strategies to reduce the climate effects on rice cultivation. The specific objectives were to: (1) simulate the potential climate impacts of rice yield for the near term (2010 2039) and the mid century (2040 2069) under two climate scenarios (RCPs 4.5 and 8.5) and (2) assess the effectiveness of the adaptation strategie s that could help reduce the yield effects of the climate change on rice cultivation for the future climate periods under the two RCPs. These two specific objectives constituted two chapters of the thesis.

PAGE 81

81 Chapter 2 The specific objectives of this chapter were to simulate the potential climate trend for the near term (2010 2049) and the midcentury (2050 2069) climate periods under the two climate scenarios (RCPs 4.5 and 8.5), to evaluate the impacts of these changes on rice productivity, and to determine th e main climate factors influencing the decrease of rice yield in the Artibonite Valley of Haiti. Results indicated that the anthesis, the physiological maturity and the yield were well calibrated and evaluated using CERES RICE The Relative Root Mean Squar e Error (RRMSE) for each of the three variables was less than 10% in both the calibration and evaluation of the model with the two cultivars (TCS10 and CAP). The minimum and the maximum temperatures were predicted to increase and the rainfall to decrease i n all three growing seasons (spring summer, summer autumn and winter spring) under both climate scenarios. The solar radiation was predicted to increase in the spring summer and the summer autumn seasons and decrease in the winter spring for both the mid t erm and the mid century. For both of these climate periods, the rainfall increased only in the spring summer season and decreased in the summer autumn and the winter spring seasons with all the five GCMs considered in the study. Rice phenology duration (an thesis and physiological maturity days) and yield of both cultivars w ere predicted to be affected by the changing climate in the Valley. In the spring summer, summer autumn and winter spring, respectively, rice anthesis duration was predicted to decrease b y 4.5 7.8, 3.2 6.5 and 0.2 5.3 days under the RCP 4.5, and by 4.7 9.4, 3.4 7.6, and 1.0 8.0 days under the RCP 8.5. For these three seasons correspondingly, the maturity duration decreased by 5.2 9.9, 6.1 10.8 and 0.2 7.8 days under the RCP 4.5 and reduced by 5.4 11.6, 6.8 12.4 and 0.4 11.2 days under the RCP 8.5. Rice yield in the spring summer and was predicted to decrease by 98 137 kg ha 1 for the near term and by 117 150 kg ha 1 for the mid century under two RCPs (4.5 and 8.5). Considering this same cli mate scenarios RCPs.4.5 and 8.5), the

PAGE 82

82 ensemble average rice yield in the summer autumn season decreased by 90 120 kg ha 1 for the near term and by 90 157 kg ha 1 for the mid century while increased in the winter spring season by 40 61 kg ha 1 and 20 45 kg ha 1 for the two climate periods (near term and mid century, respectively). Rice yield in the winter spring was predicted to increase and benefit from climate change compared to the other seasons. Chapter 3 The chapter objective was to simulate the applicatio n of alternative planting dates, different nitrogen application amounts and plant breeding to reduce the impacts of climate change on rice yield in the Artibonite Valley of Haiti. Results of the simulation showed changing that planting date could be a usef ul adaptation strategy to reduce the impacts of climate change on rice yield in the Artibonite Valley. Simulation results showed that transplanting rice plants 42 days earlier than the current date in each season increased the ensemble average annual yield by 81 105 kg ha 1 in the near term, and transplanting 28 days before the current transplanting dates in the mid century during each of the three seasons increased the annual rice ensemble average yield by 80 112 kg ha 1 under the both RCPs 4.5 and 8.5. B reeding new rice cultivars showed good potential to improve rice yield under the changing climate in the Artibonite Valley. Although virtual drought and heat tolerance were created, the 10% longer life cycle cultivar increased the yield up 2 4%. Also, the m odification of the yield potential traits improved the significantly the rice yield in the Artibonite Valley under both the RCP s 4.5 and 8.5. Increasing rice life cycle and increased the traits of the potential yield of the original cultivars showed good p otential to improve rice yield under the changing climate in the Artibonite Vall ey of Haiti. The change of fertilizer application amount was predicted to increase rice yield significantly in the Artibonite Valley under both the RCPs 4.5 and 8.5 compared to the

PAGE 83

83 application of the currently recommended application rate baseline (75 kg ha 1 ). The application of 21 0 kg N ha 1 was predicted to increase the ensemble average rice yield by 1 407 1839 kg ha 1 for the near term and by 1671 2027 kg ha 1 for the mid century, in the spring summer season ; the application of 150 kg ha 1 increased the ensemble average rice yield in summer autumn by 258 286 kg ha 1 and 375 705 kg ha 1 for the near term and the mid century, respectively. In the winter spring s eason, the application of 130 kg ha 1 increased the ensemble mean rice yield by 628 723 kg ha 1 and 543 599 kg ha 1 for the near term and the mid century, respectively.

PAGE 84

84 LIST OF REFERENCES Adams R., S., Cockshull E., K., & Cave J., C. R. (2001). Effect of Temperature on the Growth and Development of Tomato Fruits. Annals of Botany 88 (5), 869 877. https://doi.org/10.1006/anbo.2001.1524 [doi] (2013). Application of the CSM CERES Rice model for evaluation of plant density and irrigation management of transplant ed rice for an irrigated semiarid environment. Irrigation Science 31 (3), 491 506. https://doi.org/10.1007/s00271 012 0324 6 Ainsworth, E. A. (2008). Rice production in a changing climate: a meta analysis of responses to elevated carbon dioxide and elevate d ozone concentration. Global Change Biology 14 (7), 1642 1650. https://doi.org/10.1111/j.1365 2486.2008.01594.x Antle M., J., & Stckle O., C. (2017). Climate Impacts on Agriculture: Insights from Agronomic Economic Analysis. Review of Environmental Econo mics and Policy. 11 (2), 299 318. https://doi.org/10.1093/reep/rex012 [doi] (2013). Uncertainty in simulating wheat yields under climate change. Nature Climate C hange 3 827. Retrieved from https://doi.org/10.1038/nclimate1916 Asseng, S., Fillery, I. R. P., Anderson, G. C., Dolling, P. J., Dunin, F. X., & Keating, B. A. ) leaching for a de ep sand. Australian Journal of Agricultural Research 49 (3), 363 377. https://doi.org/10.1071/A97095 Babel, M. S., Agarwal, A., Swain, D. K., & Herath, S. (2011a). Evaluation of climate change impacts and adaptation measures for rice cultivation in Northea st Thailand. Climate Research 46 (2), 137 146. https://doi.org/10.3354/cr00978 Babel, M. S., Agarwal, A., Swain, D. K., & Herath, S. (2011b). Evaluation of climate change impacts and adaptation measures for rice cultivation in Northeast Thailand. Climate R esearch 46 (2), 137 146. https://doi.org/10.3354/cr00978 Bai, J., Chen, X., Dobermann, A., Yang, H., Cassman, K. G., & Zhang, F. (2010). Evaluation of NASA Satellite and Model Derived Weather Data for Simulation of Maize Yield Potential in China, 102 (1), 9 16. https://doi.org/10.2134/agronj2009.0085 Bao, Y., Hoogenboom, G., McClendon, R., & Vellidis, G. (2017). A comparison of the performance of the CSM CERES Maize and EPIC models using maize variety trial data. Agricultural Systems https://doi.org///doi. org/10.1016/j.agsy.2016.10.006 Basak, J. K., Ali, M. A., Islam, N., & Rashid, A. (2010). Assessment of the effect of climate change on boro rice production in Bangladesh using DSSAT model.

PAGE 85

85 Basso, B., Liu, L., & Ritchie, J. T. (2016). A Comprehensive Review of the CERES Wheat, Maize and Advances in Agronomy Academic Press. https://doi.org///doi.org/10.1016/bs.agron.2015.11.004 Battisti, R., Bender, F. D., & Sentelhas, P. C. (2018). Assessment of different gr idded weather data for soybean yield simulations in Brazil. Theoretical and Applied Climatology https://doi.org/10.1007/s00704 018 2383 y Beven, K., & Binley, A. (1992). The future of distributed models: Model calibration and uncertainty prediction. Hydro logical Processes 6 (3), 279 298. https://doi.org/10.1002/hyp.3360060305 Bhuvaneswari, K., Geethalakshmi, V., Lakshmanan, A., Anbhazhagan, R., & Sekhar, D. N. U. (2014). Climate change impact assessment and developing adaptation strategies for rice crop in western zone of Tamil Nadu. Journal of Agrometeorology 16 (1), 38 43. Blanc, E., & Schlenker, W. (2017). The Use of Panel Models in Assessments of Climate Impacts on Agriculture. Review of Environmental Economics and Policy. 11 (2), 258 279. https://doi.o rg/10.1093/reep/rex016 [doi] Bocchiola, D. (2015). Impact of potential climate change on crop yield and water footprint of rice in the Po valley of Italy. Agricultural Systems https://doi.org///doi.org/10.1016/j.agsy.2015.07.009 Boonwichai, S., Shrestha, S., Babel, M. S., Weesakul, S ., & Datta, A. (2019). Evaluation of climate change impacts and adaptation strategies on rainfed rice production in Songkhram River Basin, Thailand. Science of The Total Environment https://doi.org///doi.org/10.1016/j.scitotenv.2018.10.201 Boote, K. J. (2 011). Improving Soybean Cultivars for Adaptation to Climate Change and Climate Variability. Crop Adaptation to Climate Change https://doi.org/10.1002/9780470960929.ch26 Kesner, J. L. (2015). Estimation des cots des impacts du changement climatique en Hati New York: PNUD https://scholar.google.com/scholar_lookup?title=Estimation des couts des impacts du changement climatiq ue en Haiti&author=A. Borde&author=M. Huber&author=A. Goburdhun&author=A. Guidoux&author=E. Revoyron&author=E. Nsimba&author=JA. Louis&author=A. Donjia& Bryan, E., Deressa, T. T., Gbetibouo, G. A., & Ringler, C. (2009). Adaptation to climate change in Ethiopia and South Africa: options and constraints. Environmental Science & Policy https://doi.org///doi.org/10.1016/j.envsci.2008.11.002 Challinor, A. J., Watson, J., Lobell, D. B., Howden, S. M., Smith, D. R., & Chhetri, N. (2014). A meta analysis of crop yield under climate change and adaptation. Nature Climate Change 4 287. Retrieved from http://dx.doi.org/10.1038/nclimate2153

PAGE 86

86 Chun, J. A., Li, S., Wang, Q., Lee, W. S., Lee, E. Assessing rice productivity and adaptation strategies for Southeast Asia under climate change through multi scale crop modeling. Agricultural Systems https://doi.org///doi.org/10.1016/j.agsy.2 015.12.001 CNSA. (2014). Evaluation Previsionnelle de La Performance Des Rcoltes de La Campagne au Prince, Hati: MARNDR Coordination Nationale de la Securite Alimentaire (CNSA) Retrieved from http://www.cnsa509.org/Web /Etudes/Evaluation previsionnelle des recoltes campagne printemps 2014 version definitive.pdf Targeted News Service Retrieved from https://search.proquest.com/docview/1761624228 Cohe n, M., & Singh, B. (2014). Climate Change Resilience: The case of Haiti Policy File Oxfam. Retrieved from https://search.proquest.com/docview/1820775182 Constantin, J., Beaudoin, N., Launay, M., Duval, J., & Mary, B. (2012). Long term nitrogen dynamics i n various catch crop scenarios: Test and simulations with STICS model in a temperate climate. Agriculture, Ecosystems & Environment https://doi.org///doi.org/10.1016/j.agee.2011.06.006 Coucheney, E., Buis, S., Launay, M., Constantin, J., Mary, B., Garca de Cortzar Lonard, J. (2015). Accuracy, robustness and behavior of the STICS soil crop model for plant, water and nitrogen outputs: Evaluation over a wide range of agro environmental conditions in France. Environmental Modelling & Software https://doi.org///doi.org/10.1016/j.envsoft.2014.11.024 Dass, A., Nain, A. S., Sudhishri, S., & Chandra, S. (2012). Simulation of maturity duration and productivity of two rice varieties under system of rice intensification using DSSAT v 4.5/CERES Rice mo del. Journal of Agrometeorology 14 (1), 26 30. Ding, Y., Wang, W., Song, R., Shao, Q., Jiao, X., & Xing, W. (2017). Modeling spatial and temporal variability of the impact of climate change on rice irrigation water requirements in the middle and lower reac hes of the Yangtze River, China. Agricultural Water Management https://doi.org///doi.org/10.1016/j.agwat.2017.08.008 Dixit, P. N., Telleria, R., Al Khatib, A. N., & Allouzi, S. F. (2018). Decadal analysis of impact of future climate on wheat production in dry Mediterranean environment: A case of Jordan. Science of The Total Environment https://doi.org///doi.org/10.1016/ j.scitotenv.2017.07.270 Dos Santos, M. G., De Faria, R. T., Palaretti, L. F., Dantas, G. D. F., Dalri, A. B., & Lopes, A. D. S. (2016). Calibration and Testing of Cs Cropgro Model for Common Beans. Engenharia Agricola 36 (6), 1239 1249. https://doi.org/10. 1590/1809 4430 Eng.Agric.v36n6p1239 1249/2016

PAGE 87

87 Perceptions of Agricultural Risks and Risk Management Strategies. Agriculture https://doi.org/10.3390/agriculture9010010 Fan zo, J., Davis, C., McLaren, R., & Choufani, J. (2018). The effect of climate change across food systems: Implications for nutrition outcomes. Global Food Security https://doi.org///doi.org/10.1016/j.gfs.2018.06.001 Food and Agriculture Organization of the United Nations, (FAO). (2013). Food Balance Sheet. FAOSTAT Statistics database. FAO. Retrieved from http://www.fao.org/faostat/en/#data/FBS Gruda, N., Bisbis, M., & Tanny, J. (2019). Influence of climate change on protected cultivation: Impacts and sustai nable adaptation strategies A review. Journal of Cleaner Production https://doi.org///doi.org/10.1016/j.jclepro.2019.03.210 practices of high yield and high nit rogen use efficiency in Jiangsu, China. Scientific Reports 7 (1), 2101. https://doi.org/10.1038/s41598 017 02338 3 Haile, M. G., Wossen, T., Tesfaye, K., & von Braun, J. (2017). Impact of Climate Change, Weather Extremes, and Price Risk on Global Food Supp ly. Economics of Disasters and Climate Change 1 (1), 55 75. https://doi.org/10.1007/s41885 017 0005 2 (2011). Climate Impacts on Agriculture: Implications for Crop Production, 103 (2), 351 370. https://doi.org/10.2134/agronj2010.0303 He, J., Jones, J. W., Graham, W. D., & Dukes, M. D. (2010). Influence of likelihood function choice for estimating crop model parameters using the generalized likelihood uncertai nty estimation method. Agricultural Systems https://doi.org///doi.org/10.1016/j.agsy.2010.01.006 (2017). Decision Support System for Agrotechnology Transfer ( DSSAT)  https://doi.org///DSSAT.net). DSSAT Foundation, Gainesville, Florida, USA Responses of crop yield growth to global temperature and socioeconomic change s. Scientific Reports 7 7800. https://doi.org/10.1038/s41598 017 08214 4 IPCC. (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Co re Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland. Retrieved from https://www.ipcc.ch/report/ar5/syr/

PAGE 88

88 Izaurralde, R. C., Williams, J. R., McGill, W. B., Rosenberg, N. J., & Jakas, M. C. Q. (2006). Simulating soil C dynamics w ith EPIC: Model description and testing against long term data. Ecological Modelling https://doi.org///doi.org/10.1016/j.ecolmodel.2005.07.010 Jamieson, P. D., Porter, J. R., & Wilson, D. R. (1991). A test of the computer simulation model ARCWHEAT1 on whe at crops grown in New Zealand. Field Crops Research https://doi.org///doi.org/10.1016/0378 4290(91)90040 3 Jena, U. R., Swain, D. K., Hazra, K. K., & Maiti, M. K. (2018). Effect of elevated [CO 2 ] on yield, intra plant nutrient dynamics, and grain quality of rice cultivars in eastern India. Journal of the Science of Food and Agriculture 98 (15), 5841 5852. https://doi.org/10.1002/jsfa.9135 Jgadish, S. V. K., Murty, M. V. R., & Quick, W. P. (2015). Rice responses to rising temperatures challenges, perspect ives and future directions. Plant, Cell & Environment 38 (9), 1686 1698. https://doi.org/10.1111/pce.12430 Ritchie, J. T. (2003). The DSSAT cropping system model. European Journal of Agronomy 18 (3 4), 235 265. https://doi.org/10.1016/S1161 0301(02)00107 7 Singh, U. (1998). Decision support system for agrotechnology transfer: DSSAT v3. In Gordon Y Tsuji, G. Hoogenboom, & P. K. Thornton (Eds.), Understanding Options for Agricultural Production (pp. 157 177). Dordrecht: Springer Netherlands. https://doi.org/10.1007/978 94 017 3624 4_8 (2017). Toward a new generation of agricu ltural system data, models, and knowledge products: State of agricultural systems science. Agricultural Systems https://doi.org///doi.org/10.1016/j.agsy.2016.09.021 Jones, James W, He, J., Boote, K. J., Wilkens, P., Porter, C. H., & Hu, Z. (2011). Estimat ing DSSAT Cropping System Cultivar Specific Parameters Using Bayesian Techniques; Advances in Agricultural Systems Modeling 2. In Methods of Introducing System Models into Agricultural Research (pp. 365 394). Madison, WI: American Society of Agronomy, Crop Science Society of America, Soil Science Society of America. https://doi.org/10.2134/advagricsystmodel2.c13 Jones, P. G., & Thornton, P. K. (2000). MarkSim: Software to generate daily weather data for Latin America and Africa. Agronomy Journal 92 (3), 445 453. https://doi.org/10.2134/agronj2000.923445x Kapetanaki, G., & Rosenzweig, C. (1997). Impact of climate change on maize yield in central and northern Greece: A simulation study with CERES Maize. Mitigation and Adaptation Strategies for Global Change 1 (3), 251 271. https://doi.org/10.1007/BF00517806

PAGE 89

89 Wegehenkel, M. (2015). Analysis and classification of data sets for calibration and validation of agro ecosystem models Environmental Modelling & Software https://doi.org///doi.org/10.1016/j.envsoft.2015.05.009 Kim, H. Y., Ko, J., Kang, S., & Tenhunen, J. (2013). Impacts of climate change on paddy rice yield in a temperate climate. Global Change Biology 19 (2), 548 562. https://doi.org/10.1111/gcb.12047 Kontgis, C., Schneider, A., Ozdogan, M., Kucharik, C., Tri, V. P. D., Duc, N. H., & Schatz, J. (2019). Climate change impacts on rice productivity in the Mekong River Delta. Applied Geography https://doi.org///doi.org/10. 1016/j.apgeog.2018.12.004 Krishnan, P., Swain, D. K., Chandra Bhaskar, B., Nayak, S. K., & Dash, R. N. (2007). Impact of elevated CO 2 and temperature on rice yield and methods of adaptation as evaluated by crop simulation studies. Agriculture, Ecosystems & Environment https://doi.org///doi.org/10.1016/j.agee.2007.01.019 Lamy, J. D., Lotfi, K., Louissaint, J., & Tescar, R. (2011). Effet de la fertilisation phosphate et potassique sur le rendement de la varit de riz (Oryza sativa, L.) TCS10 la valle de Bouman, B. (2015). Uncertainties in predicting rice yield by current crop models under a wide range of climatic conditions. Global Change Biology 21 (3), 1328 1341. https://doi.o rg/10.1111/gcb.12758 Li, Y., Wu, W., Ge, Q., Zhou, Y., & Xu, C. (2016a). Simulating Climate Change Impacts and Adaptive Measures for Rice Cultivation in Hunan Province, China. Journal of Applied Meteorology and Climatology 55 (6), 1359 1376. https://doi.org/10.1175/JAMC D 15 0213.1 Li, Y., Wu, W., Ge, Q., Zhou, Y., & Xu, C. (2016b). Simulating Climate Change Impacts and Adaptive Measures for Rice Cultivation in Hunan Province, China. Journal of Applied Meteorology and Climatology 55 (6), 1359 1376. https://doi.org/10.1175/JAMC D 15 0213.1 the DSSAT CERES Maize model to simulate crop yield and nitrogen cycling in fields under long term continuous maize p roduction. Nutrient Cycling in Agroecosystems 89 (3), 313 328. https://doi.org/10.1007/s10705 010 9396 y Lobell, D. B., & Gourdji, S. M. (2012). The Influence of Climate Change on Global Crop Productivity. Plant Physiology 160 (4), 1686 1697. https://doi.o rg/10.1104/pp.112.208298

PAGE 90

90 Louissaint, J., & Duvivier, P. (2003). pour la fertilisation rationnelle et conomique des terres rizicoles de la valle de Lychuk, T. E., Hill, R. L., Izaurrald e, R. C., Momen, B., & Thomson, A. M. (2017). Evaluation of climate change impacts and effectiveness of adaptation options on crop yield in the Southeastern United States. Field Crops Research https://doi.org///doi.org/10.1016/j.fcr.2017.09.020 Mahajan, G ., Bharaj, T. S., & Timsina, J. (2009). Yield and water productivity of rice as affected by time of transplanting in Punjab, India. Agricultural Water Management https://doi.org///doi.org/10.1016/j.agwat.2008.09.027 Mal, S., Singh, R. B., Huggel, C., & Gr over, A. (2018). Introducing Linkages Between Climate Change, Extreme Events, and Disaster Risk Reduction. In S. Mal, R. B. Singh, & C. Huggel (Eds.), Climate Change, Extreme Events and Disaster Risk Reduction: Towards Sustainable Development Goals (pp. 1 14). Cham: Springer International Publishing. https://doi.org/10.1007/978 3 319 56469 2_1 Maldonado, W., Valeriano, T. T. B., & de Souza Rolim, G. (2019). EVAPO: A smartphone application to estimate potential evapotranspiration using cloud gridded meteorol ogical data from NASA POWER system. Computers and Electronics in Agriculture https://doi.org///doi.org/10.1016/j.compag.2018.10.032 Maplecroft. (2016). Climate Change Vulnerability Index. Retrieved from http://www.maplecroft.com/about/news/ccvi.html MARND R. (2014). Rponse de trois varits de riz (CAP, TCS10 et L1) diffrentes doses grain et de production de biomasse Retrieved from https://agriculture.gouv.ht/view/01/IMG/pdf/rapport_essai.pdf Masui, T., Matsumoto, K., Hij M. (2011). An emission pathway for stabilization at 6 Climatic Change 109 (1), 59. https://doi.org/10.1007/s10584 011 0150 5 Matthews, R. B., Kropff, M. J., Horie, T., & Bachelet, D. (1997). Simulating the impact of climate change on rice production in Asia and evaluating options for adaptation. Agricultural Systems https://doi.org///doi.org/10.1016/S0308 521X(95)00060 I Matthews, R., & Wassmann, R. (2003). Modelling th e impacts of climate change and methane emission reductions on rice production: a review. European Journal of Agronomy https://doi.org///doi.org/10.1016/S1161 0301(03)00005 4 ME. (2013). Deuxime communication nationale sur les changements climatiques, 18 1. Retrieved from https://unfccc.int/sites/default/files/resource/htinc2.pdf

PAGE 91

91 Mendelsohn, R. O., & Massetti, E. (2017). The Use of Cross Sectional Analysis to Measure Climate Impacts on Agriculture: Theory and Evidence. Review of Environmental Economics and Policy 11 (2), 280 298. https://doi.org/10.1093/reep/rex017 Ministere de l Agriculture, des R. N. et du D. R. (2015). Situation de la fili re riz 2014 2015 Retrieved from http://agriculture.gouv.ht/statistiques_agricoles/wp content/uploads /2016/11/Situation de la filire riz 2014 15.pdf Moore, F. C., & Lobell, D. B. (2014). Adaptation potential of European agriculture in response to climate change. Nature Climate Change 4 610. Retrieved from http://dx.doi.org/10.1038/nclimate2228 Myers, S. S., Zanobetti, A., Kloog, I., Huybers, P., Leak (2014). Increasing CO 2 threatens human nutrition. Nature 510 (7503), +. https://doi.org/10.1038/nature13179 Nath, H. K., & Mandal, R. (2018). Heterogeneous Climatic Impacts on Agricultural Production: Evidence from Ri ce Yield in Assam, India. Asian Journal of Agriculture and Development 15 (1), 23 42. NET, F. (2018). Haiti staple food market fundamentals. USAID USAID Retrieved from https://reliefweb.int/sites/reliefweb.int/files/resources/Haiti MFR_final_20180326 %28 1%29.pdf Olszyk, D. M., Centeno, H. G. S., Ziska, L. H., Kern, J. S., & Matthews, R. B. (1999). Global climate change, rice productivity and methane emissions: comparison of simulated and experimental results. Agricultural and Forest Meteorology https://d oi.org///doi.org/10.1016/S0168 1923(99)00065 9 J. P. (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth A ssessment Report of the Intergovernmental Panel on Climate Change IPCC. Retrieved from http://epic.awi.de/37530/ Patricola, C. M., & Wehner, M. F. (2018). Anthropogenic influences on major tropical cyclone events. Nature 563 (7731), +. https://doi.org/10. 1038/s41586 018 0673 2 (2014). Food security and food production systems. Fifth Assessment Report of the Intergovernmental Panel on Climate Change 1 82. Pranu thi, G., & Tripathi, S. K. (2018). Assessing the climate change and its impact on rice yields of Haridwar district using PRECIS RCM data. Climatic Change 148 (1), 265 278. https://doi.org/10.1007/s10584 018 2176 4 Rankinen, K., Karvonen, T., & Butterfield, D. (2006). An application of the GLUE methodology for estimating the parameters of the INCA N model. Science of The Total Environment https://doi.org///doi.org/10.1016/j.scitotenv.2006.02.034

PAGE 92

92 Ray, D. K., Gerber, J. S., MacDonald, G. K., & West, P. C. (20 15). Climate variation explains a third of global crop yield variability. Nature Communications 6 5989. Retrieved from https://doi.org/10.1038/ncomms6989 Rgis, Y. L. (2016). Rice landscape analysis for rice fortification: Haiti World Food Programme U.S. Agriculture and Climate Change: New Results. Climatic Change 57 (1), 43 67. https://doi.org/1022103315424 Riahi, K., Rao, S., Krey, V., Cho, C A scenario of comparatively high greenhouse gas emissions. Climatic Change 109 (1 2), 33 57. https://doi.org/10.1007/s10584 011 0149 y Rinaldi, M., Losavio, N., & Flagella, Z. (2003). Evaluation and application of the OILCROP SUN model for sunflower in southern Italy. Agricultural Systems https://doi.org///doi.org/10.1016/S0308 521X(03)00030 1 Ritchie, J. T., Singh, U., Godwin, D. C., & Bowen, W. T. (1998). Cereal growth, development and yield. In Go rdon Y Tsuji, G. Hoogenboom, & P. K. Thornton (Eds.), Understanding Options for Agricultural Production (pp. 79 98). Dordrecht: Springer Netherlands. https://doi.org/10.1007/978 94 017 3624 4_5 Rosenzweig, C, Antle, J. M., Ruane, A. C., Jones, J. W., Hatfi C. Z. (2018). The Agricultural Model Intercomparison and Improvement Project (AgMIP). Protocols for AgMIP Regional Integrated Assessments Version 7.0. Retrieved from http://www.agmip.org/wp content/uploads/2018/07/A gMIP Protocols for Regional Integrated Assessment v7 0 20180218 1 ilovepdf compressed.pdf Rosenzweig, Cynthia, Iglesias, A., Yang, X. B., Epstein, P. R., & Chivian, E. (2001). Climate Change and Extreme Weather Events; Implications for Food Production, Pla nt Diseases, and Pests. Global Change and Human Health 2 (2), 90 104. https://doi.org/1015086831467 (2018). Coordinating AgMIP data and models across global a nd regional scales for 1.5C and 2.0C assessments. Philosophical Transactions. Series A, Mathematical, Physical, and Engineering Sciences 376 (2119), 20160455. https://doi.org/10.1098/rsta.2016.0455 Rosenzweig, Cynthia, Strzepek, K. M., Major, D. C., Igle sias, A., Yates, D. N., McCluskey, A., & Hillel, D. (2004). Water resources for agriculture in a changing climate: international case studies. Global Environmental Change https://doi.org///doi.org/10.1016/j.gloenvcha.2004.09.003 Rosenzweig, Cynthia, & Tub iello, F. N. (2007). Adaptation and mitigation strategies in agriculture: an analysis of potential synergies. Mitigation and Adaptation Strategies for Global Change 12 (5), 855 873. https://doi.org/10.1007/s11027 007 9103 8

PAGE 93

93 Sarker, M. A. R., Alam, K., & Go w, J. (2014). Assessing the effects of climate change on rice yields: An econometric investigation using Bangladeshi panel data. Economic Analysis and Policy https://doi.org///doi.org/10.1016/j.eap.2014.11.004 Shi, P., Tang, L., Lin, C., Liu, L., Wang, H. Cao, W., & Zhu, Y. (2015). Modeling the effects of post anthesis heat stress on rice phenology. Field Crops Research https://doi.org///doi.org/10.1016/j.fcr.2015.02.023 Shrestha, S., Deb, P., & Bui, T. T. T. (2016). Adaptation strategies for rice cultiv ation under climate change in Central Vietnam. Mitigation and Adaptation Strategies for Global Change 21 (1), 15 37. https://doi.org/10.1007/s11027 014 9567 2 Singh, P., Boote, K. J., Kadiyala, M. D. M., Nedumaran, S., Gupta, S. K., Srinivas, K., & Bantila n, M. C. S. (2017). An assessment of yield gains under climate change due to genetic modification of pearl millet. Science of The Total Environment https://doi.org///doi.org/10.1016/j.scitotenv.2017.06.002 (2016). Rice (Oryza sativa L.) yield gap using the CERES rice model of climate variability for differe nt agroclimatic zones of India. Current Science 110 (3), 405 413. https://doi.org/10.18520/cs/v110/i3/405 413 Bantilan, M. C. S. (2014). Quantifying potential benefits of drought and heat tolerance in rainy season sorghum for adapting to climate change. Agricultural and Forest Meteorology https://doi.org///doi.org/10.1016/j.agrformet.2013.10.012 Stackhouse, P. W., Zhang, T., Westberg, D., Barnett, A. J., Brist ow, T., Macpherson, B., & Hoell, J. M. (2018). POWER Release 8 (with GIS Applications) Methodology Retrieved from https://power.larc.nasa.gov/documents/POWER_Data_v8_methodology.pdf Sun, M., Zhang, X., Huo, Z., Feng, S., Huang, G., & Mao, X. (2016). Uncertainty and sensitivity assessments of an agricultural hydrological model (RZWQM2) using the GLUE method. Journal of Hydrology 534 19 30. https://doi.org/10.1016/j.jhydrol.2015.12.04 5 Tao, F., & Zhang, Z. (2010). Adaptation of maize production to climate change in North China Plain: Quantify the relative contributions of adaptation options. European Journal of Agronomy https://doi.org///doi.org/10.1016/j.eja.2010.04.002 Thomson, A. M (2011). RCP4.5: a pathway for stabilization of radiative forcing by 2100. Climatic Change 109 (1 2), 77 94. https://doi.org/10.1007/s10584 011 0151 4 Tiepolo, M., & Bacci, M (2017). Tracking Climate Change Vulnerability at Municipal Level in Rural Haiti Using Open Data. In M. Tiepolo, A. Pezzoli, & V. Tarchiani (Eds.), Renewing Local Planning to Face Climate Change in the Tropics (pp. 103 131). Cham: Springer International P ublishing. https://doi.org/10.1007/978 3 319 59096 7_6

PAGE 94

94 Timsina, J., & Humphreys, E. (2006). Performance of CERES Rice and CERES Wheat models in rice wheat systems: A review. Agricultural Systems https://doi.org///doi.org/10.1016/j.agsy.2005.11.007 Tingem, M., & Rivington, M. (2009). Adaptation for crop agriculture to climate change in Cameroon: Turning on the heat. Mitigation and Adaptation Strategies for Global Change 14 (2), 153 168. https://doi.org/10.1007/s11027 008 9156 3 Tsuji, G Y. (1998). Network m anagement and information dissemination for agrotechnology transfer. In Gordon Y Tsuji, G. Hoogenboom, & P. K. Thornton (Eds.), Understanding Options for Agricultural Production (pp. 367 381). Dordrecht: Springer Netherlands. https://doi.org/10.1007/978 94 017 3624 4_18 Tubiello, F. N., Rosenzweig, C., Goldberg, R. A., Jagtap, S., & Jones, J. W. (2002). Effects of climate change on US crop production: simulation results using two different GCM scenarios. Part I: Wheat, potato, maize, and citrus. Clim Res 2 0 (3), 259 270. Retrieved from https://www.int res.com/abstracts/cr/v20/n3/p259 270/ USAID. (2011). Haiti Climate Change Gender Action Plan, 70. Retrieved from https://www.climatelinks.org/resources/haiti climate change gender action plan ccgap report USAID (2013). Climate Vulnerability Profile: Haiti USAID Retrieved from https://www.climatelinks.org/resources/haiti climate vulnerability profile USAID. (2017). Climate Risk Profile: Haiti. Retrieved from https://www.climatelinks.org/resources/climate risk profile haiti Usui, Y., Sakai, H., Tokida, T., Nakamura, H., Nakagawa, H., & Hasegawa, T. (2016). Rice grain yield and quality responses to free air CO 2 enrichment combined with soil and water warming. Global Change Biology 22 (3), 1256 1270. https://doi.o rg/10.1111/gcb.13128 Vaghefi, N., Shamsudin, M. N., Radam, A., & Rahim, K. (2013). Modelling the Impact of Climate Change on Rice Production: An Overview (Vol. 13). https://doi.org/10.3923/jas.2013.5649.5660 V an Ittersum, M. K., Cassman, K. G., Grassini, P., Wolf, J., Tittonell, P., & Hochman, Z. (2013). Yield gap analysis with local to global relevance A review. Field Crops Research https://doi.org///doi.org/10.1016/j.fcr.2012.09.009 V an Oldenborgh, G. J., va Cullen, H. (2017). Attribution of extreme rainfall from Hurricane Harvey, August 2017. Environmental Research Letters 12 (12), 124009. https://doi.org/10.1088/1748 9326/aa9ef2

PAGE 95

95 V an Vuuren, Ruijven, B. (2011). RCP2.6: exploring the possibility to keep global mean temperature increase below 2C. Climatic Change 109 (1), 95. https://doi.org/10.1007/s10584 011 0152 3 Vilayvong, S., Banterng, P., Patanothai, A., & Pannangpetch, K. (2015). CSM CERES Rice model to determine management strategies for lowland rice production. Scientia Agricola 72 (3), 229 236. https://doi.org/10.1590/0103 9016 2013 0380 Walsh, K. J. E., McBride, J. L., Klotzbach, P. J., Balachandran, S., Camargo, S. J., Holland, G., Wiley Interdisciplinary Reviews: Climate Change 7 (1), 65 89. https://doi.org/10.1002/wcc.371 Wang, D. R., Bunce, J. A., Tomecek, M. B., Gealy, D., McClung, A., McCouch, S. R., & Ziska, L. H. (2016). Evidence for divergence of response in Indica, Japonica, and wild rice to high CO 2 temperature interaction. Global Change Biology 22 (7), 2620 2632. https://doi.org/10. 1111/gcb.13279 Wang, Z., Qi, Z., Xue, L., Bukovsky, M., & Helmers, M. J. (2015). Modeling the impacts of climate change on nitrogen losses and crop yield in a subsurface drained field. Climatic Change 129 (1 2), 323 335. https://doi.org/10.1007/s10584 015 1342 1 Warszawski, L., Frieler, K., Huber, V., Piontek, F., Serdeczny, O., & Schewe, J. (2014). The Inter Sectoral Impact Model Intercomparison Project (ISI MIP): Project framework. Proceedings of the National Academy of Sciences 111 (9), 3228 3232. https: //doi.org/10.1073/pnas.1312330110 Welch, J. R., Vincent, J. R., Auffhammer, M., Moya, P. F., Dobermann, A., & Dawe, D. (2010). Rice yields in tropical/subtropical Asia exhibit large but opposing sensitivities to minimum and maximum temperatures. Proc Natl Acad Sci USA 107 (33), 14562. https://doi.org/10.1073/pnas.1001222107 White, J. W., Hoogenboom, G., Kimball, B. A., & Wall, G. W. (2011). Methodologies for simulating impacts of climate change on crop production. Field Crops Research 124 (3), 357 368. http s://doi.org/10.1016/j.fcr.2011.07.001 Wilcock, C. D., & Jean Pierre, F. (2012). Haiti Rice Value Chain assessment: Rapid diagnosis and implications for program design Oxfam America Oxfam America Research Backgrounder series. Retrieved from https://policy practice.oxfamamerica.org/publications/haiti rice value chain assessment/ Worou, O. N., Gaiser, T., Saito, K., Goldbach, H., & Ewert, F. (2012). Simulation of soil water dynamics and rice crop growth as affected by bunding and fertilizer app lication in inland valley systems of West Africa. Agriculture, Ecosystems & Environment https://doi.org///doi.org/10.1016/j.agee.2012.07.018

PAGE 96

96 Wunder, S., Noack, F., & Angelsen, A. (2018). Climate, crops, and forests: a pan tropical analysis of household in come generation. Environment and Development Economics 23 (3), 279 297. https://doi.org/10.1017/S1355770X18000116 Xiong, W., Conway, D., Lin, E., & Holman, I. (2009). Potential impacts of climate change and duction. Climate Research 40 (1), 23 35. https://doi.org/10.3354/cr00802 Xiong, W., Holman, I., Conway, D., Lin, E., & Li, Y. (2008). A crop model cross calibration for use in regional climate impacts studies. Ecological Modelling https://doi.org///doi.org/10.1016/j.ecolmodel.2008.01.005 Xu, C. C., Wu, W. X., Ge, Q. S., Zhou, Y., Lin, Y. M., & Li, Y. M. (2017). Simulating climate change impacts and potential adaptations on rice yields in the Sichuan Basin, China. Mitigation and Adaptation Strategies for Global Change 22 (4), 565 594. https://doi.org/10.1007/s11027 015 9688 2 Ya ng, J., & Zhang, J. (2006). Grain filling of cereals under soil drying. New Phytologist 169 (2), 223 236. https://doi.org/10.1111/j.1469 8137.2005.01597.x Yao, F., Xu, Y., Lin, E., Yokozawa, M., & Zhang, J. (2007). Assessing the impacts of climate change o n rice yields in the main rice areas of China. Climatic Change 80 (3), 395 409. https://doi.org/10.1007/s10584 006 9122 6 Zhang, H., Zhou, G., Liu, D. L., Wang, B., Xiao, D., & He, L. (2019). Climate associated rice yield change in the Northeast China Plai n: A simulation analysis based on CMIP5 multi model ensemble projection. Science of The Total Environment https://doi.org///doi.org/10.1016/j.scitotenv.2019.01.415 Zhang, P., Zhang, J., & Chen, M. (2017). Economic impacts of climate change on agriculture: The importance of additional climatic variables other than temperature and precipitation. Journal of Environmental Economics and Management https://doi.org///doi.org/10.1016/j.jeem.2016.12.001

PAGE 97

97 BIOGRAPHICAL SKETCH Floyid Nicolas was born and raise d in Petit Goave (Haiti). He began his education in Pierre Mendes France Kindergarten school and attended the Sacr Coeur middle school for his primary education. He went to the Lyc e Faustin Soulouque of Petit Goave which was one of the top secondary school in Petit Goave. He developed a keen interest in mathematics, physics and chemistry. In 2008, he was admitted to both the college of Medicine and the college of agriculture of the University of Haiti. He chose the College of Agriculture, where he earned a a gricultural e ngineering in 2013. After his undergraduate degree, he worked for the several NGOs private agency and the Ministry of Agriculture of Haiti as ca rtographer and GIS analyst and engineer. In 201 5 greenhouse in Minnesota and went back to Haiti the same to work as site engineer and GIS officer for the ministry of agriculture for two years. In 2017, h e was awarded a scholarship to the D epartment of Agricultural and Biological Engineering of the University of Florida where he worked on crop modeling and climate change under the guidance of Dr. Kati W. Migliaccio He graduated with his MS in 2019. Floyid will join the University of California Davis (UCD) to pursue his Ph.D. in Agricultural and Biological Engineering starting in September 2019.