i ii Matching cereal and legume crop varieties to production environments in Northeast Nigeria using Decision Support Tools (DST) H.A. Ajeigbe, A.Y. Kamara, F.M. Akinseye, P.K. Silwal, O. Faleti, A.I. Tofa, N. Kamai, J. Bebeley, and R. Solomon (Editors) 2024 iii Published by the International Institute of Tropical Agriculture (IITA), 2024 International address: IITA, Grosvenor House, 125 High Street Croydon CR0 9XP, UK Headquarters: PMB 5320, Oyo Road Ibadan, Oyo State ISBN 978-978-131-425-4 Printed in Nigeria by IITA Citation: H.A. Ajeigbe, A.Y. Kamara, F.M. Akinseye, P.K. Silwal, O. Faleti, A.I. Tofa, N. Kamai, J. Bebeley, and R. Solomon (Editors). 2024. Matching cereal and legume crop varieties to production environments in Northeast Nigeria using Decision Support Tools (DST). International Institute of Tropical Agriculture (IITA). 92 pp. Disclaimer: Mention of any proprietary product or commercial applications does not constitute an endorsement or a recommendation for its use by IITA. iv Contents Foreword ............................................................................................................................................ vi Acknowledgment .............................................................................................................................. vii Abbreviations and Acronyms .......................................................................................................... viii Chapter 1 .................................................................................................................................... 1 Introduction ................................................................................................................................ 1 References ........................................................................................................................................... 3 Chapter 2 .................................................................................................................................... 4 Project sites, soils, and weather ................................................................................................. 4 Abstract ............................................................................................................................................... 4 2.1. Project Sites ................................................................................................................................. 5 2.2. Soil Fertility Assessment ............................................................................................................. 8 2.2.2 Results ................................................................................................................................................... 8 2.3 Weather ....................................................................................................................................... 11 2.3.1. Methodology ...................................................................................................................................... 11 2.3.2 Results ................................................................................................................................................. 11 References ......................................................................................................................................... 16 Chapter 3 .................................................................................................................................. 17 Using the DSSAT Model to simulate the performance of maize, cowpea, and soybean ........ 17 Abstract ............................................................................................................................................. 17 3.1. Introduction ................................................................................................................................ 18 3.2. Calibration, evaluation and application of the Ceres-Maize Model in DSSAT to simulate performance of maize in selected communities in Adamawa and Borno States .............................. 21 3.2.1. Methodology ...................................................................................................................................... 21 3.2.2. Results ................................................................................................................................................ 23 3.3. Calibration, evaluation and application of the CROPGRO model in DSSAT to simulate performance of cowpea varieties in selected communities in Adamawa and Borno States ............. 32 3.3.1. Methodology ...................................................................................................................................... 32 3.3.2. Results ................................................................................................................................................ 33 3.4. Calibration, evaluation and application of the CROPGRO-Soybean model in DSSAT to simulate performance of soybean varieties in selected communities in Adamawa and Borno States .......................................................................................................................................................... 38 3.4.1. Methodology ...................................................................................................................................... 38 3.4.2. Results ................................................................................................................................................ 39 3.5. Conclusions ................................................................................................................................ 47 References ......................................................................................................................................... 48 Chapter 4 .................................................................................................................................. 50 Simulating the performance of sorghum, pearl millet and groundnut cultivars in diverse agroecologies of North-Eastern, Nigeria: Application of APSIM models .............................. 50 Abstract ............................................................................................................................................. 50 v 4.1. Introduction ................................................................................................................................ 50 4.2. Calibration, evaluation and application of the APSIM model to simulate the performance of sorghum varieties in selected communities in Adamawa and Borno States ..................................... 52 4.2.1. Methodology ...................................................................................................................................... 52 4.2.1.2. Model Validation (experiments and procedure of validation ..................................................... 53 4.2.2. Results ................................................................................................................................................ 54 4.3. Calibration, evaluation and application of the APSIM model to simulate the performance of millet varieties in selected communities in Adamawa and Borno States .......................................... 62 4.3.1. Methodology ...................................................................................................................................... 62 4.3.2. Results ................................................................................................................................................ 63 4.4. Calibration, evaluation and application of the APSIM model to simulate the performance of groundnut varieties in selected communities in Adamawa and Borno States ................................... 71 4.4.1 Methodology ....................................................................................................................................... 71 4.4.2. Results ................................................................................................................................................ 72 4.5. Conclusions ................................................................................................................................ 79 References ......................................................................................................................................... 79 Appendix ........................................................................................................................................... 82 vi Foreword The violence perpetrated by religious extremist and later by bandits and kidnappers have devastated much of North-East Nigeria during the past 12 years. These terrorist threats are impeding civil progress, restricting agricultural activities and resulting in a major displacement of local populations. At the same time, agriculture is facing several constraints including climate change, drought, poor soils, pest and diseases, weak economic infrastructure and markets, and paucity of progressive policies that support agricultural development. The resultant low agricultural productivity has led to alarming rates of food and nutritional insecurity, too limited livelihood opportunities, chronic underemployment, and severe malnutrition. However, improved technologies, practices and innovations are available to address these agricultural constraints. The Feed the Future Nigeria Integrated Agriculture Activity (IAA) issued under the US Government’s Global Food Security Act was awarded by USAID Nigeria to IITA and its partners on 19th July 2019 as part of USAID’s contributions to the economic recovery in the North-East, in the aftermath of the on-going insurgency in the region. IAA supports vulnerable populations to engage in basic farming activities that will improve food security, increase agricultural incomes, and improve resilience among smallholder farmers and their families in Adamawa and Borno states. IAA works with a coalition of public and private sector partners to facilitate improved agro-inputs and extension advisory services to serve vulnerable populations, strengthen the institutions that form the market system and the networks that serve smallholder farmers who have been disenfranchised by conflict, and facilitate the engagement of youth and women in commercial agribusiness activities. IITA being a science-based organisation uses science based and proven market-oriented tools starting from identification of climate resilient and market-oriented varieties of component crops. The book “Matching cereal and legume crop varieties to production environments in northeast Nigeria using decision support tools (DST)” reports on the simulation of the performance of the widely grown improved varieties of cereals and legumes using two sets of decision support tools; Decision Support Tools for Agricultural Technology Transfer (DSSAT) model and the Agricultural Production simulation model (APSIM) to recommend those that are most suitable to the agroecology in which the project works (Adamawa and Borno States). These are then deployed with improved agronomic practices and new technologies by the Activity to achieve great results. This book is intended to guide farmers, extension personnel, students of agriculture in higher institutions, researchers, and other development projects on the improved varieties of legumes and cereals to use or promote in Northeast Nigeria especially in Adamawa and Borno States to increase productivity. Kenton Dashiell Deputy Director General, Partnerships for Delivery/General Directorate International Institute of Tropical Agriculture (iita.org) file://///nas2/data/klopez/AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/Content.Outlook/AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/Content.Outlook/AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/Content.Outlook/AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/Content.Outlook/AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/Content.Outlook/FGKNP1ID/www.iita.org vii Acknowledgment We acknowledge the many people who have contributed to the development of this book other than the listed authors, especially the Feed the Future (FtF) Nigeria Integrated Agriculture Activity team for their tireless efforts and immense contribution towards the achievement of the Activity’s objectives. The Officials of the Adamawa and Borno States ADPs as well as the field and Extension Agents, were very helpful during the fieldwork that involved digging in several sites and farms. Most importantly we are grateful to the nice and hardworking farmers of these states for not only cooperating with us to dig profile pits in their farms but on most occasions also assisting us in the work as well as providing valuable information about the land use history. We thank the Management of IITA led by the Director General, Dr. Simeon Ehui, Dr. Kenton Dashiell, Dr. Alfred Dickson, Dr. Robert Asiedu, Dr. Gbassey Tarawali, and others who provided valuable advice and contributions to the project development and implementation. We also appreciate colleagues of IITA and ICRISAT at Kano Station for their assistance. Finally, we thank Dr. Ayoade Adetoye, the Activity’s AOR and his other colleagues at USAID (past and present) who have provided their active support in terms of providing technical guidance in making sure that we follow USAID rules and regulations, and the documents are of quality. This publication is a production of the Feed the Future Nigeria Integrated Agriculture Activity implemented in targeted locations of Adamawa, Borno, Gombe and Yobe states, Nigeria between 2019 and 2024, and was made possible through financial support from the United States Agency for International Development (USAID). The views expressed in this publication are those of the authors and do not necessarily reflect the views or policies of the United States Agency for International Development (USAID) or the United States Government. The designations employed and the presentation of material in this information product do not imply the expression of any opinion whatsoever on the part of the USAID concerning the legal or development status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. The mention of specific companies or products of manufacturers, whether or not these have been patented, does not imply that these have been endorsed or recommended by USAID in preference to others of a similar nature that are not mentioned. The views expressed in this information product are those of the author(s) and do not necessarily reflect the views or policies of USAID. viii Abbreviations and Acronyms ABU Ahmadu Bello University ABCOA Audu Bako College of Agriculture ADP Agricultural Development Project APSIM Agricultural Production simulation model DSSAT Decision Support Tools for Agricultural Technology Transfer FAO Food and Agriculture Organization FAOSTAT Food and Agriculture Organization and Statistical Database FMARD Federal Ministry of Agriculture and Rural Development IAA Integrated Agriculture Activity ICRISAT International Crops Research Institute for the Semi-Arid Tropics IITA International Institute of Tropical Agriculture NAERLS National Agricultural Extension Research and Liaison Services USAID United States Agency for International Development 1 Chapter 1 Introduction Maize, millet, and sorghum are the most important cereal crops in northern Nigeria (Ajeigbe et al., 2018, Kamara al., 2020). These cereal crops provide the calorie needed for the households in this region. In addition to the cereals, legume crops such as cowpea, groundnut and soybean are important components of the farming systems. They are cheap sources of food and feed because of their high protein content. Groundnut and soybean are also important cash crops because of their use in the processing industry for oil and animal feeds. Nigeria is the largest producer of millet, sorghum, and cowpea and second largest producer of maize, groundnut in Africa (FAOSTAT, 2020). Nigeria produces 2.61 million tons of cowpea, 2.89 million tons of groundnut, 10.15 million tons of maize, 2.24 million ton of millet, 6.86 million tons of sorghum, and 0.76 million tons of soybean in 2018. All the cereal and legume crops produced in northern Nigeria are also produced in northeast Nigeria because of the availability of diverse growing environments covering the Guinea, Sudan and the Sahel savannas. Despite the importance of these crops in Nigeria, yields are low compared to other countries. According to FAO, yields obtained on farmers’ fields are 913, 991, 2092, 801, 1120, and 971 kg/ha for cowpea, groundnut, maize, millet, sorghum and soybean, respectively (FAOSTAT, 2018, 2020). The yields of cereals and legumes are limited by several constraints in the Nigeria savannas. In the northeast Nigeria, poor soil fertility, intermittent drought, infestation of crop lands by parasitic weeds, and pest and diseases significantly reduce crop yields. For example, yield loss of up to 60- 80% are reported due to low plant nutrients and drought (Kamara et al., 2013). If these stresses occur together with pest and disease attacks, total yield loss of the crops will occur. Several agronomic technologies have been developed to address the effects of these biotic and abiotic constraints. For example, several Striga-resistant and drought-tolerant cowpea and maize varieties (Kamara et al, 2013; Menkir et al., 2016; Omoigui et al., 2017). Striga and drought-tolerant varieties of millet (Ajeigbe et al., 2019) and sorghum varieties (Ajeigbe et al., 2018) have been developed. Moreover, soil and crop management practices have been developed for rapid dissemination in the Nigeria savannas along with the improved crop varieties (Adnan et al., 2017, Ajeigbe et al., 2019, Kamara et al.,2016, 2009, Akinseye et al., 2020, Tofa et al., 2020). Several field trials have been carried out in the Nigerian savanna over the past 20 years to evaluate the performance of the various crop production technologies to improve the productivity of cereal and legume crops. To be able to adequately assess the performance of these technologies across the Nigeria savannas would require the establishment of several field experiments under the different environmental conditions. Moreover, there is a need to widely disseminate the crop varieties and the management technologies in northeast Nigeria where the Feed-Future project Integrated Agriculture Activity is being implemented. To be able to this, will require widespread testing across the region to identify the most suitable technologies. However, there are inherent factors limiting the quantity of field experiments that can be conducted under different soil types and climate conditions in this region, including economic and time constraints. The soils in the region are heterogenous and the weather is very variable which makes it impossible to extrapolate results from one location to another. Thus, crop simulation models represent a complementary approach to further investigate the potential impacts of crop varieties and management practices on grain yields of cereals and legumes across a range of environments. 2 Crop models are increasingly used as a tool to explore the spatio-temporal impacts of different management scenarios following calibrations at field experiments, particularly for upscaling the impacts of crop varieties and management practices from field to watershed and regional scales (Keating et al., 2003). At a larger scale, prediction models may help farmers understand how to implement the most efficient management practices for a certain genotype in a certain environment. In this booklet, we report on the simulation of the performance of the widely grown cereals and legumes in the project areas in Adamawa and Borno States using two sets of decision support tools: Decision Support Tools for Agricultural Technology Transfer (DSSAT) model and the Agricultural Production simulation model (APSIM). The booklet is divided into four sections. Section 1 provides general introduction into the problems of crop production in northern Nigeria, progress in developing technological solutions to address these constraints, and the need to use Decision Support Tools to evaluate and target the technologies to specific domains in the project areas. Section 2 describes the project sites, soils and climatic conditions. Section 3 addresses the simulation of the performance of maize, cowpea, and soybean in the selected communities in targeted Local Governments in Adamawa and Borno States using the DSSAT model. Section 4 reports on the results of the simulation of the performance of groundnut, millet, and sorghum using the APSIM model. 3 References Adnan, A.A., Jibrin M.J., Kamara, A.Y., Abdulrahman, B.L., Shuaibu A.S., and Garba, I.I. (2017). CERES-Maize Model for Determining the Optimum Planting Dates of Early Maturing Maize Varieties in Northern Nigeria. Frontiers in Plant Science, 8: 1118. Ajeigbe H.A., Akinseye, F.M., Ayuba, K., and Jonah, J. (2018). Productivity and Water Use Efficiency of Sorghum [Sorghum bicolor (L.) Moench] Grown under Different Nitrogen Applications in Sudan Savanna Zone, Nigeria. International Journal of Agronomy, vol 2018, Article ID 7676058, 11pp. https://doi.org/10.1155/2018/7676058 Ajeigbe, H.A., Akinseye, F.M., Kunihya, A., Abdullahi, A.I., Kamara A.Y. (2019). Response of pearl millet (Pennisetum glaucum, L.) to plant population in the semi-arid environments of Nigeria. Net J Agric Sci, 7(1): 13-22. FAOSTAT (2018). Food and Agricultural Organization of the United Nations (FAO), FAO Statistical Database, from http://faostat.fao.org. FAOSTAT (2020). Food and Agricultural Organization of the United Nations (FAO), FAO Statistical Database, from http://faostat.fao.org. Akinseye, F.M., Ajeigbe, H.A., Kamara A.Y., Adefisan, E.A., and Whitbread, A.M. (2020): Understanding the response of sorghum cultivars to nitrogen applications in the semi-arid Nigeria using the agricultural production systems simulator. Journal of Plant Nutrition, article: https://doi.org/10.1080/01904167.2020.1711943 Kamara A.Y., Ekeleme F., Menkir A., Chikoye D., and Omoigui L.O. (2009). Influence of nitrogen fertilization on the performance of early and late maturing maize cultivars under natural infestation with Striga hermonthica. Archives of Agronomy and Soil Science, 55(2): 125-145. Kamara, A.Y., Tofa, A.I., Boahen, S.K., Solomon, R., Ajeigbe, H.A., and Kamai, N. (2016). Effects of plant density on the performance of cowpea in Nigerian savannas. Experimental Agriculture p. 1 of 13 Cambridge University Press 2016 doi:10.1017/S0014479716000715. Kamara, A.Y., Ewansiha, S.U., and Menkir, A. (2013). Assessment of nitrogen uptake and utilization in drought tolerant and Striga resistant tropical maize varieties. Archives of Agronomy and Soil Science, 60(2): 195-206. Keating, B.A., Carberry P.S., Hammer, G.L., Probert, M.E., Robertson, M.J., Holzworth, D., Huth, N.I., Hargreaves, J.N.G., Meinke, H., and Hochman, Z. (2003). An overview of APSIM, a model designed for farming systems simulation. European Journal of Agronomy, 18, 267- 288. Menkir, A., Crossa, J., Meseka, S., Bossey, B., Ado, S.G., Obengantiwi, K., Yallou, C.G., Coulibaly, N., Olaoye, G., and Alidu, H. (2016). Comparative performance of top-cross maize hybrids under managed drought stress and variable rainfed environments. Euphytica, 212: 455–472. Omoigui. L.O., Kamara, A.Y., Alunyo, G.I., Bello, L.L., Oluoch, M., Timko, M.P. and Boukar, O. (2017). Identification of new sources of resistance to Striga gesnerioides in cowpea Vigna unguiculate accessions. Genetic Resources and Crop Evolution 64(5): 901-911. Tofa, A.I., Chiezey, U.F., Babaji, B.A., Kamara, A.Y., Adnan, A.A., Beah, A., Adam, A.M. (2020). Modeling Planting-Date Effects on Intermediate-Maturing Maize in Contrasting Environments in the Nigerian Savanna: An Application of DSSAT Model. Agronomy, 10, 871. https://www.hindawi.com/79361942/ https://www.hindawi.com/93902126/ https://www.hindawi.com/10243563/ https://www.hindawi.com/15152976/ http://faostat.fao.org/ 4 Chapter 2 Project sites, soils, and weather Abstract The performances of the cereal (maize, sorghum, millet and rice) and legume (cowpea, soybean and groundnut) varieties were simulated in 19 communities across 10 Local Government Areas (LGAs) in Adamawa State and in 15 communities across 5 LGAs in Borno State. These covered the northern Guinea savanna (NGS) zone, southern Guinea savanna (SGS) and Sudan Savanna (SS) zones in the two states. For soil characterization and soil sampling, profile pits were dug in the selected sites in both States (Adamawa and Borno). The profiles and soil types were classified using the FAO guidelines. Soil samples collected were shipped to IITA, Ibadan, Nigeria for analysis. All laboratory analyses were carried out at the Analytical Services Laboratory of IITA. Long-term weather data was sourced from gridded downscaled Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) for daily rainfall (Funk et al., 2015) and National Aeronautics and Space Administration (NASA) database for Climatology Resource for Agroclimatology http://power.larc.nasa.gov/that include minimum and maximum air temperature and solar radiation. Thereafter, the two datasets were merged using R scripts were developed to append CHIRPS and NASA power data together and convert each location into a format readily ingestible by the APSIM model for the 33 selected sites. For the long-term simulation, the soil parameters used for both models (DISSAT and APSIM) were obtained from on-site soil characterization using geospatially buffering points in at least 20 km radius using ArcGIS map of the reference indicating the sites/LGAs. The results of the weather data and soil analysis are given below. Thess were used to for the simulation exercise. 5 2.1. Project Sites The performance of the cereal and legume varieties was simulated in 19 communities across 10 Local Government Areas (LGAs) in Adamawa State and in 15 communities across 5 LGAs in Borno State (Table 1 and Fig. 1). In Adamawa most of the LGAs are in the northern Guinea savanna (NGS) zone except Fufore in Yola south, Yalwa Dembore in Yola north and Nassarawo-Demsa in Demsa LGA, which lie in the southern Guinea savanna (SGS) and Dulmava in Hong and Guyaku in Gombi LGA which lie in the Sudan Savanna (SS) zones. In Borno States, most of the targeted LGAs are in Sudan savanna except Gwaskara, Kubo and Lakundum in Shani LGA which are in the northern Guinea savanna zone. In the SGS, temperature varies annually and seasonally over the zone with average maximum temperature in the growing season within the range of 26-28oC whereas minimum temperature ranges between 18-22 oC (Omotosho et al. 2013; Ayanlade, 2016). Rainfall distribution in the zone is unimodal. Average annual rainfall range between 1000 mm to 1524 mm and spread over 181-210 days which defines the growing season (Jagtap, 1995; Ayanlade, 2016). The soils in this zone have been identified mainly as Lithosols, Ferralic combisols, Feric acrisols, Oxic haplustalfs and Luvisols (FAO/UNESCO, 1974). In the NGS, the length of growing period is between 151-180 days (Jagtap, 1995). It has a unimodal rainfall distribution averaging between from 900 to 1000 mm annually, and maximum temperatures varied from 28 to 40°C (Atehnkeng et al., 2008). According to world reference base FAO classification, the dominant soil in the NGS are Luvisols (FAO, 2006). The Sudan savanna is characterized by high annual temperature (28-32 °C), short growing season around 90 days and low rainfall ranging from 600 to 800 mm (Adnan et al., 2017). The soil of the Sudan savanna is sandy and porous, with rapid drainage of water. The dominant soil types mainly found in the zone are Alfisols, and Entisols according to world reference base FAO classification. 6 Table 2.1 Summary of the project sites that were used for simulation of crop performance. S/No State _LGA _Location AEZ Code Longit. Lat. 1 Adamawa Demsa Mbula Kuli NGS DMK 12.301568 9.457453 2 Demsa Nassarawo Demsa SGS DNS 12.150069 9.296248 3 Girei Wuroshi NGS GJB 12.616352 9.468659 4 Girei Daneyel NGS GIT 12.513956 9.547608 5 Gombi Tawa NGS GOT 12.685600 10.169090 6 Gombi Guyaku SS GOG 12.663390 10.345880 7 Guyuk Chikila NGS GUC 11.971910 9.772365 8 Guyuk Lakumna NGS GUG 11.989722 9.920833 9 Hong Dulmava SS HOB 12.982394 10.301400 10 Hong Hushere Zum NGS HOH 13.080656 10.103753 11 Numan Bare NGS NB 12.110769 9.584298 12 Numan Kikan_Kodomti NGS NK 11.987783 9.460814 13 Shellenge Jonkolo - Lama NGS SHEG 12.177973 9.899652 14 Sheleng Lakati_Libbo/ NGS SHEWY 12.250196 9.695414 15 Song Sabon Gari NGS SOSG 12.593541 9.840488 16 Song Suktu NGS SOS 12.424821 9.637458 17 Yola North Yelwa -Jambore SGS YNY 12.504630 9.261650 18 Yola South Fufure SGS YSNG 12.650420 9.173600 1 Borno Bayo Balbaya SS BABL 11.764809 10.584837 2 Bayo Briyel SS BABR 11.649672 10.371014 3 Bayo Jara-Dali SS BAJD 11.731594 10.275863 4 Biu Buratai SS BIB 12.415800 10.767500 5 Biu Kabura SS BIK 12.265300 10.739200 6 Biu Mathau SS BIM 12.109700 10.721400 7 Biu Tum SS BIT 12.488100 10.822800 8 Hawul Kwajaffa SS HAK 12.483106 10.516721 9 Hawul Puba Vidau SS HAPV 12.187900 10.522375 10 Hawul Sakwa Hema SS HASH 12.389373 10.386722 11 Kwayakusar Bila Gusi NGS KKBG 12.047606 10.519175 12 Kwayakusar Kurbo Gayi SS KKKG 11.957516 10.384040 13 Shani Gwaskara NGS SHAG 12.158012 10.227146 14 Shani Kubo NGS SHAK 12.085300 10.140000 15 Shani Lakundum SS SHAL 12.050556 10.055556 LGA = Local Government Area, AEZ=Agro-ecological zone 7 Fig. 2.1. Map showing Study areas in Adamawa and borno states 8 2.2. Soil Fertility Assessment 2.2.1 Methodology For the long-term simulation, the soil parameters used for the both models were obtained from on- site soil characterization using geospatially-buffering points in at least 20 km radius using ArcGIS map of the reference indicating the sites/LGAs. For soil characterization and soil sampling, profile pits were dug in the selected sites in both States (Adamawa and Borno). The profiles and soil types were classified using the FAO guidelines (FAO, 2006). Soil samples collected were shipped to IITA, Ibadan, Nigeria for analysis. All laboratory analyses were carried out at the Analytical Services Laboratory of IITA. Total soil organic carbon (total C) was measured using a modified Walkley and Black chromic acid wet chemical oxidation and spectrophotometric method (Heanes, 1984). Total nitrogen (total N) was determined using a micro-Kjeldahl digestion method (Bremner, 1996). Soil pH in water (S/W ratio of 1:2.5) was measured using a glass electrode pH meter and the particle size distribution following the hydrometer method (Gee and Or, 2002). Available phosphorus was extracted using the Bray 1 method (Bray and Kurtz, 1945). Phosphorus in the extract was determined colorimetrically by the molydo-phosphoric blue method using ascorbic acid as a reducing agent. K was analyzed based on Mehlich 3 extraction procedure (Mehlich, 1984). 2.2.2 Results According to the soil analysis, most of the topsoils in Adamawa State were coarse textured with higher sand contents. Out of the 18 study sites, 72% had sandy loam, 17% clay, and 11% sandy clay loam texture (Table 2). The soil pH for the communities in Adamawa State ranged from 5.9 (Jonkolo-Lama in Shelleng) to 9.0 (Fufore in Yola South). More than 55% of the soils had slightly acidic (6.1–6.5) to neutral (7.3–7.8) soil reactions. The soil organic carbon (OC) contents in the State ranged from 0.22% in Daneyel in Girei LGA and Suktu in Song LGA to 0.90% in Guyuk area. The distribution of soil in the study areas showed that most of the soils (67%) in the State had low (0.4 – 1.0%) OC levels. The total soil N contents in the soils ranged from very low (< 0.05%) to low (0.06-0.1%) and 67% of the study locations in State fell within the very low N fertility class. The soil available P varied among the locations with very low (< 3.0 mg kg-1) at Woroshi in Girei, Tawa in Gombi, Chikila in Guyuk, Lakumna in Guyuk, Dulmava in Hong, Hushere-Zum in Hong, Jonkolo-Lama in Shelleng, Sabon-Gari in Song and Yelwa-Jambore in Yola North LGA; low available P (3 - 7 mg kg-1) was found in Demsa-Nassarawa in Demsa LGA, Bare in Numan, Lakati-Libbo in Shelleng and Suktu in Song LGA, while high P (> 20 mg kg-1) contents were found in Mbula kuli in Demsa LGA, Kikan_Kodomti in Numan and Fufore in Yola South LGA. This indicated that, 50% of the study locations in State fell within the very low P fertility class. Exchangeable K levels were moderate (0.3 cmol+ kg-1) to high (> 0.3 cmol+ kg-1) in 78% of the locations. Table 3 shows the summary of topsoil properties of pedons used for model applications in Borno State. Majority of the subsurface soils were also coarse textured with higher sand contents, out of the 15 study sites 47% had sandy loam, 27% clay, and 26% s loamy sand texture. The soil pH for the communities in Borno State ranged from 6.1 (Balbaya) to 8.4 (Briyel) in Bayo LGA. More than 70% of the soils had slightly acidic (6.6 – 7.2) to slightly alkaline (7.3 – 7.8) soil reactions. The soil OC contents in the State ranged from 0.12% at Mathau to 0.78% at Kabura in Biu area. Eight (8) communities equivalent to 53% of the study area had very low OC (< 0.4%) level. The total soil N contents in the soils ranged from very low to low fertility status with very low (< 0.05%) status found in Balbaya, Bila Gusi, Briyel, Buratai, Gwaskara, Jara-Dali, Kubo, Kurba, Mathau, Puba Vidau, Sakwa-shema and Tum communities, while Kabura, Kwajaffa and Lakundum communities fell within the low (0.06-0.1%) N fertility class. With the exception of 9 Gwaskara in Shani LGA the topsoil available P at all the locations in Borno State fell within very low (< 3.0 mg kg-1) fertility class. This indicated that, 93% of the study locations in State fell within the very low P fertility class. Exchangeable K levels were 7% low (< 0.15 cmol+ kg-1); 33% moderate (0.16 - 0.3 cmol+ kg-1); 60% high (> 0.3 cmol+ kg-1) in the State. Table 2.2 Subsurface physical and chemical properties used for model applications in Adamawa State. Location Profile depth BD OC Sand Silt Clay pH N Meh. P K LGA Community (cm) (g/cm3) (%) (%) (%) (%) (in H20) (%) (ppm) cmol/kg Demsa Mbula-Kuli 0-20 1.76 0.84 59 23 18 7.8 0.06 32.1 0.5 Demsa Nassarawo-Demsa 24-180 2.18 0.66 65 15 20 8.3 0.06 3.8 0.89 Girei Daneyel 31-200 1.76 0.22 81 7 12 7.0 0.01 10.9 0.3 Girei Woroshi 14-94 2.16 0.54 65 19 16 6.4 0.04 1.17 0.36 Gombi Guyaku 19-120 1.7 0.35 79 9 12 6.6 0.03 2.14 0.22 Gombi Tawa 15-127 1.79 0.62 75 13 12 6.7 0.05 3.38 0.21 Guyuk Chikila 30-180 2.18 0.90 15 19 66 8.5 0.08 2.55 0.13 Guyuk Lakumna 20-200 1.77 0.90 25 23 52 7.3 0.10 1.59 0.65 Hong Dulmava 27-201 1.82 0.51 67 15 18 7.5 0.06 1.03 0.17 Hong Hushere-Zum 41-205 1.93 0.46 80 8 12 6.3 0.03 2.41 0.40 Numan Bare 25-200 1.62 0.35 74 9 17 6.6 0.02 4.07 0.20 Numan Kikan_Kodomti 22-200 1.76 0.66 71 9 20 7.3 0.04 13.7 0.20 Shelleng Lakati-Libbo 27-200 1.83 0.30 78 9 13 7.4 0.01 5.04 0.20 Shelleng Jonkolo-Lama 15-200 2.06 0.33 78 10 12 5.9 0.02 0.89 0.14 Song Sabon-Gari 31-200 1.73 0.66 25 33 42 6.2 0.04 1.45 0.4 Song Suktu 35-210 2.08 0.22 71 11 18 6.3 0.03 6.56 0.20 Yola North Yelwa-Jambore 24-155 2.19 0.4 77 11 12 6.5 0.03 1.8 0.09 Yola South Fufore 20-145 1.98 0.54 65 17 18 9.0 0.02 32.1 0.10 BD=bulk density, OC=organic carbon content, N= percentage of Nitrogen and P=Available Phosphorus 10 Table 2.3. Subsurface physical and chemical properties used for model applications in Borno State. Location Profile depth BD OC Sand Silt Clay pH N Meh. P K LGA Community (cm) (g/cm3) (%) (%) (%) (%) (H20) (%) (ppm) cmol/kg Bayo Balbaya 9-200 1.59 0.29 83 7 10 6.1 0.01 1.03 0 Bayo Briyel 15-200 1.32 0.39 19 29 52 8.4 0.02 2.69 0.4 Bayo Jara-Dali 8-200 1.55 0.33 51 13 36 6.6 0.02 1.72 0.3 Biu Buratai 29-150 1.63 0.17 74 8 18 7.6 0.02 2.69 0.6 Biu Kabura, 22-101 1.36 0.78 36 38 26 7.1 0.06 0.89 9 Biu Mathau 12.0-94 1.62 0.12 90 0 10 7.4 0 2.83 0.8 Biu Tum 12-200 1.4 0.19 28 24 48 7.4 0.01 1.17 0.6 Hawul Kwajaffa 30-110 1.31 0.54 16 27 57 7.4 0.06 2.28 0.7 Hawul Puba Vidau 10-200 1.32 0.4 18 19 63 8.3 0.02 0.89 0.6 Hawul Sakwa Hema 15-170 1.57 0.52 74 9 17 7 0.04 0.76 0.1 Kwayakusar Bila Gusi 80-200 1.59 0.48 67 15 18 6.5 0.02 2.14 0.1 Kwayakusar Kurba Gayi 10-200 1.6 0.32 75 9 16 7.2 0.01 1.03 0.1 Shani Gwaskara 19-200 1.57 0.34 72 13 15 7.1 0.01 11.5 0.1 Shani Kubo 33-200 1.54 o.46 64 13 23 7.3 0.02 1.31 0.8 Shani Lakundum 16-200 1.52 0.73 72 10 18 7.3 0.07 13.6 9 BD=bulk density, OC=organic carbon content, N= percentage of Nitrogen and P=Available Phosphorus 11 2.3 Weather 2.3.1. Methodology Long-term weather data was sourced from gridded downscaled Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) for daily rainfall (Funk et al., 2015) and National Aeronautics and Space Administration (NASA) database for Climatology Resource for Agroclimatology http://power.larc.nasa.gov/ that include minimum and maximum air temperature and solar radiation. CHIRPS produced satellite-based rainfall products with relatively high resolutions (5.5 km) and quasi-global coverage (50 oS- 50 oN) for daily, pentadal, and monthly precipitation. The data/parameters in NASA power are provided on a global grid with a spatial resolution of 0.5° latitude by 0.5° longitude. Thereafter, the two datasets were merged using R scripts were developed to append CHIRPS and NASA power data together and convert each location into a format readily ingestible by the APSIM model for the 33 selected sites. 2.3.2 Results The long-term climatic condition of the selected communities/LGAs in both States is typical of the savannah agroecologies with three seasons, a hot and humid season from June to October during which crops are cultivated, a dry and cool season from November to February, and a dry and hot season from March to May (Dingkuhn et al., 2008). The long-term (1985-2017) rainfall indicates the rainy season starts in May and ends in October with the highest peak observed in the month August (Table 4 and 5). The results further reveal about 50 - 60% of seasonal rainfall were observed in the month of July and August and indicates high inter-seasonal variability (CV) ranging from 18 to 23 %. All the sites showed a distinct mono-modal rainfall pattern and warming temperature throughout the year. However, Fig. 2 and 3 showed that maximum temperature was faster decreasing into the growing season than minimum temperature. Also, there was no significant inter-annual variability observed among the sites for both temperatures, but maximum temperature indicated higher values (CV) varing from 3.0 to 3.7% than minimum temperature ranged from 2.0 to 2.3 in both states. In Adamawa State, the annual seasonal rainfall for most sites over the 33-year period (1985-2017) ranged from 868–893 mm, meanwhile Dulmava, and Hushere Zum in Hong LGA, and Guyaku and Tawa, Gombi LGA observed higher seasonal rainfall between 1042 and 1104 mm (Table 4). The average monthly maximum temperature across the sites over the climatic period ranged between 27.5 and 39.1 oC (Fig. 2a), while average monthly minimum temperature ranged from 15.8 to 24.9 oC (Fig. 2b). Similarly, in Borno State, the annual seasonal rainfall for most sites over the 33-year period (1985-2017) ranged from 883–998 mm (Table 5). Average monthly maximum temperature across the sites over the climatic period ranged between 27.8 and 38.9 oC (Fig. 3a) while average monthly minimum temperature ranged from 15.5 to 24.7 oC (Fig. 3b). For minimum temperature trend, the lowest value is observed in January, which coincided with a dry and cool season between November and February, while the highest value was observed in April indicating the hottest period of the year. The lowest maximum temperature was observed in August which coincided with the peak of rainy season in both states while the highest maximum temperature was observed in March coincided with the hottest month of the year occur between March and May. 12 Table 2.4 Variability analyses of monthly and seasonal rainfall in the simulation sites in Adamawa State from 1985 to 2017. Annual-total seasonal rainfall from May-Oct; Stdev- Standard deviation from mean; CV- coefficient of variation in percentage LGA Community May Jun Jul Aug Sep Oct Annual Stdev C.V (%) Demsa Demsa- Nassarawo 102.1 121.2 189.3 234.3 172.7 73.5 893 188 21 Demsa Mbula Kuli 95.9 115.7 186.5 225.8 168.1 58.6 851 181 21 Girei Daneyel 99.8 118.1 202.9 240.5 156.9 54.4 873 191 22 Girie Woroshi 103.3 126.4 216.5 244 156.4 55.7 902 191 21 Gombi Guyaku 117.9 155.9 228.9 308.8 176.6 99.1 1087 230 21 Gombi Tawa 134.2 149.6 237.1 293.3 192.4 97.2 1104 239 22 Guyuk Lakumna 91.8 110.3 167.5 258.2 174.9 68.9 872 185 21 Guyuk Chikila 98.5 106.5 178.4 249.7 165.2 67.8 866 186 21 Hong ushereZum 120 133.8 211.7 266.5 196.7 113 1042 241 23 Hong Dulmava 109.9 150.6 225.5 302.8 202.2 113.1 1104 247 22 Numan Bare 91.9 107.4 176.9 244.2 162.9 80.6 864 194 22 Numan Kodomti 91.1 109.5 176.8 243.2 170.2 75 866 194 22 Shelleng Lakati-Libbo 95.2 109.6 186.8 250.2 155.2 74.9 872 191 22 Shelleng Jonkolo- Lama 97.6 115 182.4 268.6 166.2 73.1 903 197 22 Song Sabon-Gari 99.8 119.5 211.3 269.7 181.8 82.1 964 212 22 Song Suktu 99.6 116.3 211.4 256.5 157.8 61.5 903 199 22 Yola North Yelwa- Jambore 102.1 125.4 206.6 218 163.5 52.2 868 189 22 Yola South Fufore 103.8 140.6 220.6 218.5 160.5 51.4 895 190 21 13 Fig. 2.2a &b: Average monthly variations of Maximum and Minimum temperatures between 1985 and 2017 across the simulation sites in Adamawa State. The coefficients of variation (CV) ranged from 3.0- 3.7% for maximum temperature and 2.0-2.3% for minimum temperature. 14 Table 2.5 Variability analyses of monthly and seasonal rainfall in the simulation sites in Borno State from 1985 to 2017. LGA Community May Jun Jul Aug Sep Oct Annual Stdev C.V (%) Bayo Balbaya 87.9 141.3 202.9 287.9 167.4 67.4 955 206 22 Bayo Briyel 93.2 129 174.2 242.7 182.7 61.1 883 182 21 Bayo Jara-Dali 78.4 136.8 202.8 289 204.4 80.3 992 217 21 Biu Kabura 72.5 142.4 209.7 316.1 149.3 48.4 939 188 20 Biu Mathau 78.3 144.4 204.4 312.1 165.6 51.9 957 174 18 Biu Tum 86.2 149.8 218.1 317.4 170 56.9 998 204 20 Biu Buratai 77.4 144.3 210.9 318.4 148.5 45.6 945 191 20 Hawul Kwajaffa 99.7 142.3 204.3 306.7 179.3 51.2 983 186 19 Hawul Puba Vidau 96.6 144.2 199.6 299.8 188.3 60.3 989 191 19 Hawul Sakwa Hema 93.3 144.2 206.9 307.4 176.8 60.2 989 186 19 KwayaKusar Bila-Gusi 98.9 124.5 190.6 268.6 183.4 75.7 942 189 20 Kwayakusar Kurba Gayi 85.5 145.9 213.1 303.1 166.2 61.1 975 199 20 Shani Gwaskara 83.5 142.1 198.5 295.4 201.6 74.9 996 192 19 Shani Kubo 97.3 121.6 181.9 262.2 192.2 72.3 927 186 20 Shani Lakundum 85.2 146 220.2 307.1 158.2 77.9 995 213 20 Annual-total seasonal rainfall from May-Oct; Stdev- Standard deviation from mean; CV- coefficient of variation in percentage 15 Fig. 2.3a &b Average monthly variations of Maximum and Minimum temperatures between 1985 and 2017 across the simulation sites in Borno State. The coefficients of variation (CV) ranged from 3.0-3.7% for maximum temperature and 2.0-2.3% for minimum temperature. 16 References Adnan, A.A., Jibrin M.J., Kamara, A.Y. Abdulrahman, B.L., Shuaibu A.S., and Garba, I.I. (2017). CERES-Maize Model for Determining the Optimum Planting Dates of Early Maturing Maize Varieties in Northern Nigeria. Frontiers in Plant Science, 8: 1118. Atehnkeng, J., Ojiambo, P.S., Donner, M., Ikotun, T., Sikora, R.A., Cotty, P.J., and Bandyopadhyay, R. (2008). Distribution and toxigenicity of Aspergillus species isolated from maize kernels from three agro‐ecological zones in Nigeria. International Journal of Food and Microbiology, 122: 74–84. Ayanlade, A. (2016). Seasonality in the daytime and night-time intensity of land surface temperature in a tropical city area. Science of the total Environment, 557, pp. 415-424. Doi: 10.1016/j.scitotenv.2016.03.027. Bray, R.H., and Kurtz, L.T. (1945). Determination of total, organic, and available forms of phosphorus in soils. Soil Science, 59: 39-45. Bremmer, J.M. (1996). Nitrogen-Total. In: D. L. Sparks. Method of Soil Analysis Part 3 - Chemical Methods, SSSA Book Series 5, Madison, Wisconsin, USA. Pp. 1085 - 1122. Dingkuhn, M., Kouressy, M., Vaksmann, M., Clerget, B., Chantereau, J., 2008. A model of sorghum photoperiodism using the concept of threshold-lowering during prolonged appetence. European Journal of Agronomy, 28, 74–89. FAO (2006). Base Reference Base for Soil Resources. A framework for international classification, correlation and communication. World Soil Resources Reports 103, FAO, Rome. 145 pp. FAO–UNESCO (1974). Soil map of the world. Vol. I-legend. Paris. 59 pp. Funk, C., Peterson, P., Landsfeld, M. et al. (2015). The climate hazards infrared precipitation with stations—a new environmental record for monitoring extremes. Sci Data 2, 150066. https://doi.org/10.1038/sdata.2015.66. Gee, G.W., and Or, D., (2002). Particle-size analysis. Methods of Soil Analysis Part 4. Physical Methods SSSA Book Series, 255-293. doi:10.2136/sssabookser5.4.c12. Heanes, D.L. (1984). Determination of total organic-C in soils by an improved chromic acid digestion and spectrophotometric procedure. Commun. Soil Sci. Plant Anal. 15:1191-1213. Jagtap, S.S., (1995). Changes in Annual, Seasonal and Monthly Rainfall in Nigeria during 1961- 1990 and Consequences to Agriculture. Discovery and Innovation, 7 (4): 311-426. Mehlich, A. (1984). Mehlich 3 Soil Test Extractant: A Modification of Mehlich 2 Extractant. Comm. Soil Sci. Plant Anal. 15 (12): 1409–1416. Omotosho, J. B., Agele, S.O., Balogun, I.A. and Adefisan, E.A. (2013). Climate variability, crop- climate modeling and water ecophysiology research: implications for plant’s capacities for stress acclimation, yield production and food security. Global Journal of Plant Ecophysiology, 3(2), 56–69. http://dx.doi.org/10.1016/j.scitotenv.2016.03.027 https://doi.org/10.1038/sdata.2015.66 17 Chapter 3 Using the DSSAT Model to simulate the performance of maize, cowpea, and soybean in Adamawa and Borno States Abstract Cropping system simulation models present an important opportunity for extrapolating short- duration field experimental results to other years and locations using long-term weather and soil information. To make recommendations for suitable crop varieties in Adamawa and Borno States, we calibrated and validated the CERES-maize, CROPGRO-soybean, and CROPGRO-cowpea models using secondary data collected from northern Nigeria. A close agreement was obtained between simulated and observed values with a low RMSE and a high d index for all measured parameters for all crops. After confirming the credibility of the three models, sensitivity analyses were carried out to test the performance of some selected improved cowpea, maize, and soybean varieties. For each crop, a 30-year sensitivity analysis was conducted in 15 communities in Borno and 18 communities in Adamawa, in northeast Nigeria, using the DSSAT model. For maize, the sensitivity analysis showed that medium-maturing and drought-tolerant (IWD C2 W and DT STR W) produced grain yields that were 20 and 25% higher than those of TZL COMP1 Syn W and 99 EVDT, respectively. The variety DT STR W produced grain yields that were 15 and 18% higher than that of TZL COMP 1 Syn W in Adamawa and Borno, respectively, while the increase was 20% over that of 99 EVDT in both locations. For soybean, the variety TGX1951-3F produced the highest grain yield in both States, while TGX1448-2E produced the lowest grain yield. The variety TGX1951-3F produced grain yields that were 20, 23, 17, and 8% higher than that of TGX1448- 2E, TGX1835-10E, TGX1987-10F, and TGX1904-3F, respectively, in Adamawa. The yields were higher by 21, 17, 13, and 9% for the same varieties in Borno. For cowpea, the simulation results showed that the medium-maturing Striga-resistant variety (IT99K-573-1-1) recorded the highest grain yield of above 1 ton ha-1 in both States. The highest grain yield of 1116 kg ha-1 was simulated at Mathau in Biu Local Government Area (LGA)while the lowest grain yield of 960 kg ha-1 was simulated at Bila Gusi in Kwayakusar LGA in Borno State. In Adamawa State, the highest grain yield of 1101 kg ha-1 was simulated at Guyaku in Gombi LGA while the lowest grain yield of 731 kg ha-1 was simulated at Yola North. The models simulated higher grain yields for all the crops in Borno than in Adamawa State. We concluded that the CERES and CROPGRO models can accurately predict the performance of grain crop varieties in the two States. The maize varieties IWD C2 W and DT STR W; soybean varieties TGX1951-3F and TGX1904-6F; and Striga- resistant cowpea varieties IT99K-573-1-1 and UAM 09 1051-1 can be recommended for production and dissemination in the study areas of the two States. 18 3.1. Introduction Maize, cowpea, and soybean are among the major and important staple and cash crops in Nigeria. Maize (Zea mays L.) is a very important cereal crop and is grown in virtually all the geo-ecological zones of Nigeria. Maize is most productive in the middle and the northern belts of Nigeria where sunshine is adequate, and rainfall is moderate. The lowland humid forest zone which is characterized by high rainfall and humidity is not particularly suitable for maize production due to the high incidence of the pest and diseases, low light intensity during the growing season and low soil fertility. The recent achievements by breeders in the development and release of superior varieties of maize with higher yield potential and better resistance to insect pests and diseases play a central role in increasing maize production in Nigeria. Nigeria is currently the 9th largest producer of maize in the world and the 2nd largest producer in Africa after South Africa. The total annual national production has increased from 658,000 MT in 1978 to about 10,155,027 MT in 2018 (FAOSTAT, 2018). Cowpea (Vigna unguiculata L. Walp) is the most important grain legume crop in Nigeria. It is widely cultivated for food and its seed is the major source of high-quality plant protein in human diet. The seed protein content ranges between 23% and 30% and contain most of the essential amino acids. Nigeria is the largest cowpea producer in the world with an annual production of 2.2 million tonnes from about 4 million hectares (FAOSTAT, 2018). Soybean (Glycine max L.) is becoming a major food and cash crop due to its high cash value and relative ease of cultivation and is widely used in the food and feed industry. Soybean is the world’s leading source of oil and protein. It has the highest protein content of all food crops and is second to groundnut in terms of oil content among food legumes. It contains high protein content and high-quality oil of about 40% and 20%, respectively. It contributes to improving soil fertility by providing biologically fixed nitrogen, increasing soil organic matter and is used in crop rotation to reduce Striga infestation on farmers’ fields. Nigerian soybean production is rising steadily spurred by favourable grower price and sustained high demand for soybean by products over the past years. Nigeria soybean domestic outputs has increased to about 758,033 MT in 2018 (FAOSTAT, 2018). Despite being one of the leading producers of cowpea, maize and soybean, Nigeria’s demand for these commodities exceeds its supply and the deficit is met by imports from neighbouring countries. With the increase in Nigeria’s population from the current approximate population of 170 million to 310 million by 2050 (UNDESA/PD, 2015), the demand is going to increase even further. This is due to the facts that the average yields for cowpea, maize and soybean are quite low compared to other developed countries and even the world average. The major limiting factors to potential production in Nigeria include climate variability (especially drought and high temperature), low soil nutrient level particularly nitrogen and phosphorus, infestation by parasitic weeds such as Striga and Alectra. Other limitations to high production include poor management practices such as inappropriate sowing time, limited use of inputs especially fertilizer and improved seeds. Therefore, improving and sustaining crop productivity is a critical need in Nigeria and this will mostly occur in the Sudan and Guinea savannah regions where yield potential is much higher than in the forest due to low solar radiation and high humidity. To address these constraints and improve crop production in Nigeria, international researchers in collaboration with national partners have developed improved crop varieties that are tolerance to drought, heat stress, pest and diseases (Menkir et al., 2006, 2007; Badu-Apraku et al., 2013, 2016; Omoigui et al., 2018; Kamara et al. 2004; Kamara, 2017). In addition, complementary agronomic management practices such as right fertilizer, optimum fertilizer rates and sowing windows have been developed and evaluated in the region to increase yields of the improved varieties on farmers’ fields. Dissemination of these technologies to farmers is being channelled through field testing and 19 demonstrations, provision of advisory services on crop management and storage, and through the organization of farmers’ field schools, field days and radio shows. A project, Integrated Agriculture Activity funded by the USAID in northeast Nigeria seeks to disseminate improved crop varieties and complementary technologies in Adamawa and Borno States. This would require information on the most suitable crop varieties for targeted Local Government Areas (LGAs) in the States. To provide such information would require testing these varieties in combination with a range of improved crop production technologies across several locations and LGAs. However, these traditional methods of technology dissemination have some limitations. Reports on the performance of technologies are largely site specific and do not take into consideration variability in soils and climate conditions outside the areas where the technologies are tested (Adnan et al. 2020, Tofa et al., 2020). To assess the performance of these technologies on a large scale would require time consuming and expensive large-scale experiments across crop growing regions like the savannas in northern Nigeria. An alternative to address these limitations is the use of crop models that simulate crop yield under different soil and climate conditions. Cropping system simulation models such as; Crop Environment Resource Synthesis (CERES) Maize, SOYGRO-soybean, and CROPGRO in Decision Support System for Agricultural Technology Transfer (DSSAT) present very important opportunity for extrapolating short-duration field experimental results to other years and other locations making use of long-term weather and soil information (Hoogenboom et al., 2017). DSSAT has been tested and evaluated extensively by many researchers across locations and found good correlations between observed and simulated values for a wide range of experimental practices against field data and environmental conditions (Banterng et al., 2010; Jibrin et al., 2012; Adnan et al. 2020; Tofa et al., 2020). To be able to make site-specific recommendations for suitable crop varieties and crop management practices in Adamawa and Borno States, we calibrated and validated the CERES-maize, SOYGRO-soybean and CROPGRO models using secondary data collected from northern Nigeria and used the results to simulate the performance of the varieties of cowpeas, maize, and soybean (Table 1). 20 Table 3.1 Source and description of crop varieties used in the study. Crop Variety Source Maturity (days) Seed Colour Characteristics Yield potential Maize 99 EVDT STR IITA 90-95 white tolerant to drought and resistant to Striga hermonthica 4.5 t/ha IWD C2 W IITA 106-110 white Tolerance to drought stress and Striga infestation 6.9 t/ha TZL COMP 1 SYN IITA 110-120 white Non-drought tolerant but Striga infestation 6.4 t/ha DT STR W IITA 95-100 white Tolerance to drought stress, low soil N and Striga infestation 5.5 t/ha Soybean TGX1835-10E IITA 80-90 cream Non shattering 1.5 t/ha TGX1987-62F IITA 80-90 cream Shatters 2.0 t/ha TGX1951-3F IITA 95-100 cream Non shattering 3.0 t/ha TGX1448-2E IITA 110-120 cream Non shattering 2.5 t/ha TGX1904-6F IITA 105-110 cream Non shattering 3.0 t/ha Cowpea IT99K-573-1-1 IITA 70-75 white Resistant to Striga 2.6 t/ha IT90K-277-2 IITA 75-85 white Susceptible to Striga 2.7 t/ha UAM09 1051-1 UAM 75-85 brown Resistant to Striga 2.0 t/ha 21 3.2. Calibration, evaluation and application of the Ceres-Maize Model in DSSAT to simulate performance of maize in selected communities in Adamawa and Borno States 3.2.1. Methodology 3.2.1.1. Model calibration The main objective of model calibration was to adapt the model parameters to local environmental conditions (e.g. soil types and weather conditions) and crop cultivars so as to gain a good overall agreement between simulated and observed values. Calibration trials for maize were established under optimum conditions in diverse locations in Northern Nigeria from 2017 to 2019 cropping seasons to generate genetic coefficients of diverse maize varieties. Four maize varieties (99 EVDT, IWD C2 W, TZL COMP1 SYN and DT STR W) were planted between 1st and 2nd week of July. The maize calibration trials were established in 3 locations; Samaru Zaria in the northern Guinea savanna, Bayero University, Kano and Audu Bako Collage of Agricultural (ABCOA), Danbatta both in the Sudan savanna ecology). The varieties included both early and medium maturing varieties based on their superior agronomic performance and adaptations to biotic and abioticfactors. Soil samples were collected from the calibration site and analysed for nutrient content. Information on weather at the experimental sites were obtained from WatchDog weather stations installed at the sites. For model calibration, the DSSAT crop model requires genotype specific parameters (GSPs), which are specific for each cultivar. GSPs allow the model to simulate the performance of diverse varieties under different soil, weather and management conditions (Hunt et al., 1993). GSPs of the maize varieties were first calibrated by adjusting the six coefficients P1, P2, P5, G2, G3, and phyllochron interval (PHINT) which describe the growth and development characteristics for each individual variety. Three parameters (P1, P2 and P5) define the life cycle development characteristics, two coefficients (G2 and G3) define growth and yield characteristics and one coefficient, PHINT, defines leaf tip appearances (Jones et al., 1986). Development coefficients are calculated in degree days (or thermal time) in the CERES‐Maize. Thermal time in any given day is equal to mean air temperature minus base temperature (Ritchie et al., 1998). GDD = 𝑇𝑚𝑎𝑥 + 𝑇𝑚𝑖𝑛 2 − 𝑇𝑏𝑎𝑠𝑒 where GDD is growing degree days, Tmax is maximum temperature, Tmin is minimum temperature and Tbase is base temperature (Tbase for maize = 8 °C). GDD is cumulative and is measured in °C day−1. In the CERES‐Maize model, the GSPs were calibrated by comparing simulated and measured data for days to anthesis, days to maturity, biomass, and grain yield from the calibration experiments. Since all the varieties are not in DSSAT, we created them in the genetic file (MZCER047.CUL) of DSSAT‐CSM. Initial values of the GSPs were obtained from the generic early season cultivar (990001 EARLY SEASON) and generic medium season cultivar (990002 MEDIUM SEASON) for the medium maturing varieties, which were already available in the genotype files. The computed crop specific parameters values for the cultivars were copied into MZCER047.CUL file to operate the simulations. The Generalized Likelihood Uncertainty Estimation (GLUE) Coefficient Estimator module (He et al., 2010) fixed in the DSSAT model was used to estimate the GSPs for both maize varieties. The soil, weather, and crop management information were used to provide the environmental calibration for the model. For model calibration, water and nitrogen balance simulation controls were switched off, to ensure that no stress for water or nitrogen were 22 simulated since near‐optimal conditions were assumed for water and nitrogen in the calibration experiments. 3.2.1.2. Model validation The calibrated coefficients were used for model validation. Validation is the making of comparison of model predictions with experimental data which have not been previously used for model development and calibration. The model was validated with field data sets collected from nitrogen experiments at various locations by comparing the observed and simulated results. The experiments were conducted in two years (2015 and 2016) and two locations (Samaru Zaria and Iburu). The data were used for model evaluation for IWD C2 W (SAMMAZ-15), DT STR W (SAMMAZ-26) and ZL Comp1 Syn (SAMMAZ-16) varieties using four levels of nitrogen (0, 60, 120 and 180 kg N ha-1) application. For variety EVDT 99 W STR (SAMMAZ-27), independent data sets collected from mize N response to 5 levels of nitrogen (0, 30, 60, 90 and 120 kg N ha-1) were used. These data were collected during 2016 and 2017 cropping seasons at Abuja and Samaru Zaria. Soil samples were collected from all experimental sites and analyzed for nutrient content. Weather data were collected from the WatchDog weather stations installed at all the sites. Model validation was done to test the parameters already optimized in the calibration exercise using independent experimental data. Information on soil and weather were used as input for model validation. Data used in model evaluation include final grain yield and shoot dry matter ha−1. Model statistics used to evaluate model performance are root mean square error (RMSE) and d- statistic (Willmott et al., 1982). RMSE =√1/𝑁 ∑(𝑂𝑖 − 𝑃𝑖)2 Where: P = predicted, O= observe, N= number of observations within each treatment. 𝑑 = 1 − ∑ (𝑚𝑖 − 𝑆𝑖) 2𝑛 𝑖=1 ∑ |𝑆�́�| 𝑛 𝑖=1 + |𝑚𝑖̀|)2 Where: 𝑆�́� = 𝑆𝑖 − �̅� and 𝑚�́� = 𝑚𝑖 − �̅� The d statistic is recommended for making cross-comparisons when the d value is both relative and has bounded measures. According to the d-statistic, the closer the index value is to one, the better the agreement between the two variables that are being compared. Easy-grapher program in DSSAT was used to graph and compare simulated model outputs with observed data and also calculate model performance statistics. 3.2.1.3. Model application (Seasonal analysis) After confirming the credibility of the model, sensitivity analysis was carried out to test the performance of some selected improve cowpea, maize and soybean varieties as presented in Table 1. The sensitivity analysis was conducted in 15 communities in Borno and 18 communities in Adamawa, in the northeast Nigeria, using the seasonal analysis tool of DSSAT. For maize scenario analysis, the sowing date was set at June 30 in all the locations at both States. A compound fertilizer (NPK 15:15:15) was used in the model to supply 60 kg each of N, P2O5 and K2O ha-1 at 14 days after sowing (DAS) as first application. Urea (46 % N) was used to supply the remaining dose (60 kg N ha-1) of nitrogen at 45 DAS. Generally, sowing was done at soil depth of 5 cm, with a sowing density of 5.3 plants per square meter. 23 3.2.2. Results 3.2.2.1. Model calibration The results for the model calibration experiments for maize varieties are presented in Figures 1-4. A close agreement was obtained between simulated and observed values for all four measured parameters. The statistical values of simulated and measured days to anthesis and physiological maturity ranged between (RMSE = 1.9 to 2.3 days) and (d-index = 0.82 to 0.97) for all varieties. The comparison between simulated and observed grain yield were also quite good for all the maize varieties. The RMSE values for grain yield ranged from 158 kg ha-1 for DT STR W to 470 kg ha- 1 for IWD C2 W while d-index values were above 0.90 for all varieties except DT STR W that has 0.61. The predictions for shoot dry yields were also good all for all varieties; the RMSE values for shoot dry yield ranged from 483 kg ha-1 for 99 EVDT to 1727 kg ha-1 for IWD C2 W while d- index value ranged was between 0.80 and 0.88 for all varieties. Generally, the coefficients of determination (R2) for the calibration of all the tested parameters for all varieties were good. y = 0.6058x + 20.358 R² = 0.6058 30 35 40 45 50 55 60 30 40 50 60 A n th e s is d a y ( d a p ) m e a s u re d Anthesis day (dap) simulated (a) RMSE =1.9 D =0.82 y = 0.6804x + 28.291 R² = 0.8902 80 85 90 95 100 80 85 90 95 100 P h y s io lo g ic a l m a tu ri ty d a y ( d a p ) m e a s u re d Physiological maturity day (dap) simulated (b) RMSE =2.2 D =0.93 y = 1.0205x - 75.644 R² = 0.813 4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000 4000 4500 5000 5500 6000 Y ie ld a t h a rv e s t m a tu ri ty ( k g /h a ) m e a s u re d Yield at harvest maturity (kg/ha) simulated (c) RMSE =172 D =0.94 y = 0.8247x + 2482.4 R² = 0.65 12000 13000 14000 15000 16000 17000 12000 14000 16000 T o p s w e ig h t a t m a tu ri ty ( k g /h a ) m e a s u re d Tops weight at maturity (kg [dm]/ha) simulated (d) RMSE =483 D =0.88 24 Fig. 3.1: Comparison of simulated and measured anthesis (a), physiological maturity (b), grain yield at maturity (c) and shoot dry weight (d) for 99 EVDT maize variety. Fig. 3.2 Comparison of simulated and measured anthesis (a), physiological maturity (b), grain yield at maturity (c) and shoot dry weight (d) for TZL Comp1 Syn W maize variety. y = 0.9205x + 5.4032 R² = 0.8236 40 45 50 55 60 65 40 45 50 55 60 65 A n th e si s d ay ( d ap ) m e as u re d Anthesis day (dap) simulated (a) RMSE =1.9 D =0.93 y = 1.0867x - 8.4915 R² = 0.8994 60 70 80 90 100 110 120 60 80 100 120 P h ys io lo gi ca l m at u ri ty d ay ( d ap ) m e as u re d Physiological maturity day (dap) simulated (b) RMSE =2 D =0.97 y = 0.9271x + 401.92 R² = 0.694 0 1000 2000 3000 4000 5000 6000 7000 0 2000 4000 6000 Y ie ld a t h ar ve st m at u ri ty ( kg [ d m ]/ h a) m e as u re d Yield at harvest maturity (kg [dm]/ha) simulated (c) RMSE =245 D =0.91 y = 0.5618x + 6222.1 R² = 0.5804 5000 7000 9000 11000 13000 15000 17000 19000 5000 10000 15000 20000 To p s w e ig h t at m at u ri ty ( kg [ d m ]/ h a) m e as u re d Tops weight at maturity (kg [dm]/ha) simulated (d) RMSE =1152 D =0.8 25 Fig. 3.3 Comparison of simulated and measured anthesis (a), physiological maturity (b), grain yield at maturity (c) and shoot dry weight (d) for DT STR W maize variety. y = 1.0917x - 6.3226 R² = 0.801 1:1 line 40 45 50 55 60 65 70 40 50 60 70 A n th e si s d ay ( d ap ) m e as u re d Anthesis day (dap) simulated (a) RMSE =2.1 D =0.93 y = 1.0112x - 0.5135 R² = 0.9047 1:1 line 80 85 90 95 100 105 110 115 120 80 90 100 110 120 P h ys io lo gi ca l m at u ri ty d ay ( d ap ) m e as u re d Physiological maturity day (dap) simulated (b) RMSE =2 D =0.97 y = 0.7808x + 1139.5 R² = 0.2215 4500 5000 5500 6000 4500 5000 5500 6000 Y ie ld a t h ar ve st m at u ri ty ( kg /h a) m e as u re d Yield at harvest maturity (kg/ha) simulated (c) RMSE =158 D =0.61 y = 0.6654x + 4796.7 R² = 0.8984 5000 7000 9000 11000 13000 15000 17000 19000 5000 10000 15000 20000 To p s w e ig h t at m at u ri ty ( kg /h a) m e as u re d Tops weight at maturity (kg/ha) simulated (e) RMSE =784 D =0.88 26 Fig. 3.4 Comparison of simulated and measured an thesis (a), physiological maturity (b), grain yield at maturity (c) and shoot dry weight (d) for IWD C2 W maize variety. y = 0.7462x + 16.277 R² = 0.9041 1:1 line 40 45 50 55 60 65 70 40 50 60 70 A n th e si s d ay ( d ap ) m e as u re d Anthesis day (dap) simulated (a) RMSE =1.9 D =0.93 y = 0.9304x + 9.1772 R² = 0.9278 1:1 line 60 70 80 90 100 110 120 60 80 100 120 P h ys io lo gi ca l m at u ri ty d ay ( d ap ) m e as u re d Physiological maturity day (dap) simulated (b) RMSE =2.3 D =0.95 y = 0.8866x + 545.52 R² = 0.7522 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0 2000 4000 6000 8000 Y ie ld a t h ar ve st m at u ri ty ( kg [ d m ]/ h a) m e as u re d Yield at harvest maturity (kg/ha) simulated (c) RMSE =470 D =0.92 y = 0.5934x + 5648.7 R² = 0.7034 5000 7000 9000 11000 13000 15000 17000 19000 5000 10000 15000 20000 To p s w e ig h t at m at u ri ty ( kg [ d m ]/ h a) m e as u re d Tops weight at maturity (kg/ha) simulated (d) RMSE =1727 D = 0.75 27 3.2.2.2. Model validation The accuracy of the CERES-Maize model simulations and performance of genetic coefficients were assessed by running the model with independent data sets collected during 2015 and 2016 seasons for IWD C2 W, DT STR W and ZL Comp1 Syn using four levels of nitrogen (0, 60, 120 and 180 kg N ha-1) application at Iburu and Zaria. For the variety 99 EVDT, independent data sets under 5 levels of nitrogen (0, 30, 60, 90 and 120 kg N ha-1) application were used. The data were collected during 2016 and 2017 cropping seasons at Abuja and Samaru Zaria. Grain yield at maturity and shoot dry yield at maturity were used for model evaluation (Tables 2-4). The model's evaluation of grain yield was good at all N treatment levels for the four varieties in each location. In all the locations and years, the model slightly underestimated or overestimated the measured parameters for all the four maize varieties at various N levels. However, the under or over estimations were within the acceptable range of below 25 %. In all the locations there was a good fit in the model prediction of grain yield with low RMSE and high d-index values. The values of RMSE for the four varieties ranged from 584 to 745 kg ha-1. In all cases, d-index values for grain yield were above 0.93 indicating that the model is robust and accurate in measuring grain yield. The overall RMSE for shoot yield ranged from 1746 to 2339 kg ha-1 with d-index values also above 0.9 for all the varieties. 2 8 T a b le 3 .2 S im u la te d ( S ), o b se rv ed ( O ) an d s im u la te d m in u s o b se rv ed ( S - O ) o f g ra in y ie ld a n d s h o o t y ie ld o f IW D C 2 W a n d D T S T R W o b ta in ed fr o m v al id at io n e x p er im en ts c o n d u ct ed a t tw o l o ca ti o n s o v er a t w o -y ea r p er io d . IW D C 2 W D T S T R W L o ca ti o n Y ea r N r at e (k g h a-1 ) G ra in y ie ld ( k g h a-1 ) S h o o t y ie ld ( k g h a-1 ) G ra in y ie ld ( k g h a-1 ) S h o o t y ie ld ( k g h a-1 ) S O S - O S O S - O S O S - O S O S - O Ib u ru 2 0 1 5 0 3 6 3 4 3 3 -7 0 2 5 2 9 2 2 2 2 3 0 7 1 1 2 1 1 0 3 6 8 5 5 2 8 4 5 5 8 1 -2 9 7 6 0 2 8 2 2 2 6 8 3 1 3 9 7 8 9 6 6 6 8 8 1 2 0 8 2 5 6 4 3 7 5 6 -1 1 9 2 8 8 0 7 9 4 5 6 -6 4 9 1 2 0 5 6 6 4 3 9 9 6 1 6 6 8 1 1 8 0 5 9 5 8 5 2 2 2 0 3 9 1 2 4 2 6 3 -3 5 1 1 1 9 7 0 1 1 3 4 5 6 2 5 1 8 0 6 6 6 4 4 8 4 4 1 8 2 0 1 2 1 9 6 1 1 4 5 4 7 4 2 5 3 8 8 4 5 6 6 8 2 2 1 4 5 9 6 1 2 6 6 9 1 9 2 7 2 0 1 6 0 2 8 4 6 1 4 -3 3 0 3 7 9 2 1 6 6 0 2 1 3 2 3 0 9 1 1 3 5 -8 2 6 1 9 7 6 4 9 7 9 -3 0 0 3 6 0 3 0 9 4 3 6 1 0 -5 1 6 8 6 3 0 5 8 5 9 2 7 7 1 2 9 3 3 3 6 2 8 -6 9 5 9 3 3 4 8 8 3 1 5 0 3 1 2 0 5 3 5 2 5 6 1 9 -2 6 7 1 4 4 8 6 8 6 5 9 5 8 2 7 5 4 6 5 5 0 3 7 4 2 8 1 4 7 8 7 1 2 2 6 1 2 5 2 6 1 8 0 6 1 2 5 5 8 5 9 2 6 6 1 5 2 9 6 1 0 3 8 1 4 9 1 5 6 3 1 1 5 9 3 1 3 8 0 1 7 4 8 3 1 4 8 6 4 2 6 1 9 S a m a ru 2 0 1 5 0 2 1 4 7 3 0 -5 1 6 4 0 8 9 1 6 1 4 2 4 7 5 2 3 3 9 9 6 -7 6 3 9 5 7 4 7 9 9 -3 8 4 2 6 0 3 0 2 8 3 6 8 0 -6 5 2 9 9 7 2 6 2 8 9 3 6 8 3 2 6 8 0 3 6 4 0 -9 6 0 8 1 2 9 1 0 6 0 1 -2 4 7 2 1 2 0 5 3 7 4 5 5 0 3 -1 2 9 1 2 3 6 3 8 9 6 4 3 3 9 9 4 4 4 9 4 8 6 8 -4 1 9 1 2 5 7 3 1 2 1 9 7 3 7 6 1 8 0 5 9 8 1 5 3 6 1 6 2 0 1 2 3 3 7 1 0 2 6 1 2 0 7 6 4 9 7 9 5 5 9 3 -6 1 4 1 4 5 5 1 1 3 8 9 3 6 5 8 2 0 1 6 0 5 0 6 7 3 4 -2 2 8 4 1 6 7 2 0 4 3 2 1 2 4 4 9 1 6 7 5 -1 8 4 1 8 1 4 3 0 2 5 -1 2 1 1 6 0 2 9 5 7 2 7 7 0 1 8 7 9 0 0 5 6 5 0 5 2 5 0 0 3 2 7 5 4 1 2 8 -8 5 3 9 4 9 6 9 4 0 2 9 4 1 2 0 5 3 7 8 5 4 0 5 -2 7 1 3 4 6 8 9 0 9 6 4 3 7 2 5 8 8 3 5 4 7 6 4 0 7 1 4 9 8 6 1 2 1 3 5 2 8 5 1 1 8 0 6 0 1 8 5 5 3 1 4 8 7 1 3 6 4 5 1 0 2 9 3 3 3 5 2 6 6 1 1 5 3 0 9 1 3 0 2 1 7 1 5 8 1 4 8 8 3 2 2 7 5 R M S E 7 0 9 2 3 3 9 7 4 5 1 9 9 0 d v a lu e 0 .9 7 0 .9 4 0 .9 6 0 .9 5 D - W il lm o tt i n d ex o f a g re em en t (W il lm o tt , 1 9 8 2 ) ra n g in g f ro m 0 t o 1 , 1 b ei n g p er fe ct a g re em en t 2 9 T a b le 3 .3 S im u la te d ( S ), o b se rv ed ( O ) an d s im u la te d m in u s o b se rv ed ( S - O ) o f g ra in y ie ld a n d s h o o t y ie ld o f T Z L C o m p 1 S y n W o b ta in ed f ro m v al id at io n e x p er im en ts c o n d u ct ed a t tw o l o ca ti o n s o v er a t w o -y ea r p er io d . T Z L C o m p 1 S y n L o ca ti o n Y ea r N r at e (k g h a-1 ) G ra in y ie ld ( k g h a-1 ) S h o o t y ie ld ( k g h a-1 ) S O S – O S O S - O Ib u ru 2 0 1 5 0 7 8 5 9 9 1 -2 0 6 4 0 2 4 5 0 5 1 -1 0 2 7 6 0 3 2 2 6 2 7 8 7 4 3 9 1 0 5 7 8 7 7 5 5 2 8 2 3 1 2 0 5 3 0 2 3 8 4 9 1 4 5 3 1 5 2 8 0 1 1 0 0 6 4 2 7 4 1 8 0 5 9 0 9 4 3 4 8 1 5 6 1 1 7 3 7 4 1 3 8 3 6 3 5 3 8 2 0 1 6 0 2 7 3 5 3 0 -2 5 7 1 8 5 8 2 5 7 9 -7 2 1 6 0 2 7 3 8 2 8 5 3 -1 1 5 8 2 0 3 6 9 8 8 1 2 1 5 1 2 0 4 4 9 4 3 7 3 0 7 6 4 1 2 4 8 5 9 6 3 7 2 8 4 8 1 8 0 5 1 2 0 4 2 3 2 8 8 8 1 4 6 3 2 1 1 5 3 7 3 0 9 5 S a m a ru 2 0 1 5 0 2 0 7 6 3 3 -4 2 6 1 7 5 0 3 7 7 2 -2 0 2 2 6 0 2 7 8 4 3 9 4 1 -1 1 5 7 8 9 0 3 9 8 5 2 -9 4 9 1 2 0 4 6 5 1 4 6 8 3 -3 2 1 3 4 2 9 1 2 2 5 5 1 1 7 4 1 8 0 5 0 3 5 4 7 5 1 2 8 4 1 4 9 3 1 1 2 7 0 7 2 2 2 4 2 0 1 6 0 6 4 9 7 1 1 -6 2 3 1 0 5 4 2 2 6 -1 1 2 1 6 0 3 1 8 4 4 0 1 3 -8 2 9 1 0 0 0 5 9 6 2 1 3 8 4 1 2 0 5 0 4 5 5 0 0 8 3 7 1 4 4 1 8 1 0 1 3 6 4 2 8 2 1 8 0 5 4 3 0 5 3 8 2 4 8 1 5 8 9 0 1 2 6 7 9 3 2 1 1 R M S E 7 3 0 2 5 1 1 d v a lu e 0 .9 6 0 .9 2 D - W il lm o tt i n d ex o f a g re em en t (W il lm o tt , 1 9 8 2 ) ra n g in g f ro m 0 t o 1 , 1 b ei n g p er fe ct a g re em en t 3 0 T a b le 3 .4 S im u la te d ( S ), o b se rv ed ( O ) an d s im u la te d m in u s o b se rv ed ( S - O ) o f g ra in y ie ld a n d s h o o t y ie ld o f 9 9 E V D T o b ta in ed f ro m v al id at io n ex p er im en ts c o n d u ct ed a t tw o l o ca ti o n s o v er a t w o -y ea r p er io d . L o ca ti o n Y ea r N r at e (k g h a-1 ) G ra in y ie ld ( k g h a-1 ) S h o o t y ie ld ( k g h a-1 ) S O S - O S O S - O A b u ja 2 0 1 6 0 4 6 4 1 2 5 3 -7 8 9 1 5 6 6 3 8 2 6 -2 2 6 0 3 0 1 8 8 3 2 3 3 6 -4 5 3 4 9 8 0 4 7 2 5 2 5 5 6 0 2 8 7 8 3 1 5 9 -2 8 1 7 3 8 7 6 2 7 9 1 1 0 8 9 0 3 4 1 9 4 0 0 3 -5 8 4 8 8 3 2 9 0 9 8 -2 6 6 1 2 0 3 5 4 9 4 1 6 5 -6 1 6 9 3 2 0 9 7 1 7 -3 9 7 2 0 1 7 0 1 0 3 5 1 4 3 5 -4 0 0 2 8 3 2 4 2 0 0 -1 3 6 8 3 0 3 2 6 0 2 4 4 8 8 1 2 7 4 3 8 6 2 3 6 1 2 0 2 6 0 4 2 2 4 3 6 0 8 6 1 6 9 5 6 8 7 1 1 2 2 4 5 6 9 0 4 3 6 5 4 1 6 1 2 0 4 9 7 6 5 8 6 3 6 1 1 2 9 1 2 0 4 2 9 0 4 3 6 1 -7 1 9 5 1 7 9 4 1 2 1 0 5 S a m a ru Z a ri a 2 0 1 6 0 4 8 3 1 3 9 1 -9 0 8 2 0 0 5 4 7 5 0 -2 7 4 5 3 0 2 0 0 9 2 4 7 5 -4 6 6 6 3 1 0 5 8 4 2 4 6 8 6 0 3 6 8 1 3 3 9 0 2 9 1 9 8 1 6 8 0 6 0 1 7 5 6 9 0 4 6 3 1 4 1 1 0 5 2 1 1 2 1 7 7 1 0 4 5 4 1 7 2 3 1 2 0 5 1 0 9 4 2 7 6 8 3 3 1 3 5 7 7 1 1 1 4 8 2 4 2 9 2 0 1 7 0 3 1 3 1 4 3 5 -1 1 2 2 1 6 9 5 4 1 3 3 -2 4 3 8 3 0 1 8 9 1 2 4 4 9 -5 5 8 5 8 9 4 5 9 3 1 -3 7 6 0 3 6 1 2 3 6 0 8 4 9 3 3 6 6 7 5 4 2 5 8 2 9 0 4 4 5 6 4 1 6 1 2 9 5 1 1 6 0 8 9 1 5 3 2 4 5 5 1 2 0 4 8 1 9 4 3 6 1 4 5 8 1 2 7 5 7 1 0 4 3 9 2 3 1 8 R M S E 5 8 4 1 7 4 6 d v a lu e 0 .9 5 0 .9 1 D - W il lm o tt i n d ex o f a g re em en t (W il lm o tt , 1 9 8 2 ) ra n g in g f ro m 0 t o 1 , 1 b ei n g p er fe ct a g re em en t 31 3.2.2.3. Model application (Seasonal analysis) Results of seasonal analysis for mean grain yields in different locations conducted by the DSSAT model over 30-year period are presented in Tables 5 and 6 for Borno and Adamawa, respectively. In the two States, the location had significant long-term effect on grain yield of all the varieties. On average, the varieties produced between 5 and 7% higher grain yield in Borno when compared with that of Adamawa State. However, the four maize varieties responded similarly in both Adamawa and Borno States. The medium maturing and drought tolerant (IWD C2 W and DT STR W) varieties produced grain yields (above 4 t/ha) that were higher than those of the medium maturing and drought sensitive (TZL COMP1 Syn W) variety and the early maturing and drought tolerant (99 EVDT) variety. In Both Adamawa and Borno, IWD C2 W produced grain yields which were 20 and 25% higher than those of TZL COMP1 Syn W and 99 EVDT, respectively. DT STR W produced grain yields that were 15 and 18% higher than that of TZL COMP 1 Syn W in Adamawa and Borno, respectively, while the increase was 20% higher than that of 99 EVDT in both locations. TZL COMP 1 Syn W and 99 EVDT did not significantly differ in grain yield in all the locations. For the two States, the two varieties IWD C2 W and DT STR W are recommended for production and dissemination. However, the yield produced by 99 EVDT is within the acceptable range for early maturing maize varieties in northern Nigeria. To adapt to varying weather conditions, this variety is also highly recommended particularly when rains are late or early cessation of rainfall is envisaged. Table 3.5 Maize grain yield (kg ha-1) results of 30 years (1985-2014) seasonal analysis in Borno, northeast Nigeria. Location IWD C2 W DT STR W TZL COMP1 99-EVDT Mean Bayo Balbaya 4619 4333 3920 3913 4196 Bayo Briyel 4749 4624 3980 3869 4306 Bayo Jara-Dali 4841 4695 4057 3983 4394 Biu Buratai 4711 4552 3937 3796 4249 Biu Tum 4782 4597 4021 3960 4340 Biu Kabura 4829 4697 4046 3930 4376 Biu Mathau 4837 4707 4048 3929 4380 Hawul Kwajaffa 1767 2051 1019 820 1414 Hawul Puba Vidau 4511 4458 3772 3647 4097 Hawul Sakwa Hema 4490 4441 3755 3624 4078 Kwayakusar Bila Gusi 4330 4151 3587 3508 3894 Kwayakusar Kurbo Gayi 4847 4707 4064 3991 4402 Shani Gwaskara 4896 4809 4080 3969 4439 Shani Kubo 4376 4230 3622 3533 3940 Shani Lakundum 4894 4807 4079 3967 4437 Mean 4498 4390 3732 3629 Table 3.6 Maize grain yield (kg ha-1) results of 30 years (1985-2014) seasonal analysis in Adamawa, northeast Nigeria. LGA Community IWD C2 W DT STR W TZL COMP1 99-EVDT Mean Demsa Mbula-kuli 4384 4220 3654 3518 3944 Demsa Nassarawo-Demsa 4204 4127 3555 3353 3810 Girei Daneyel 4131 3875 3532 3415 3738 Girei Woroshi 4478 4283 3747 3651 4040 Gombi Guyaku 4269 3956 3678 3687 3898 Gombi Tawa 4505 4406 3769 3672 4088 Guyuk Chikila 4351 4210 3599 3496 3914 Guyuk Lakumna 3521 3474 2885 2563 3111 Hong Dulmava 4501 4400 3762 3670 4083 32 Hong Garari Hushere- Zum 4503 4313 3784 3717 4079 Numan Bare 4328 4027 3647 3597 3900 Numan Kikan_Kodomti 4275 4097 3622 3478 3868 Shelleng Jonkolo-Lama 4179 3965 3471 3412 3757 Shelleng Lakati-Libbo 4343 4060 3658 3592 3913 Song Sabon-Gari 4018 3937 3336 3018 3577 Song Suktu 4096 4000 3369 3076 3635 Yola north Yelwa-Jambore 4138 3920 3446 3311 3704 Yola South Fufure 4488 4387 3704 3299 3970 Mean 4262 4092 3568 3418 3.3. Calibration, evaluation and application of the CROPGRO model in DSSAT to simulate performance of cowpea varieties in selected communities in Adamawa and Borno States 3.3.1. Methodology 3.3.1.1 Model calibration Six experiments were conducted in two sites at Bayero University Kano and ABCOA, Danbatta both in the Sudan savannas of Nigeria from 2016 to 2018 cropping seasons. Each experiment consisted of three cowpea varieties (UAM091051-1, IT99K-573-1-1 and IT90K-277-2) of varying maturity group and yield potentials. The experiments were laid out in RCBD with three replications each. These experiments were conducted under optimum management practices. Soil samples were collected from the experimental sites and analyzed and results input in the model during model calibration. Parameters measured and optimized during the calibration process included days to 50% flowering, days to maturity and final grain yield (kg ha-1) at harvest maturity. Calibration of a model is the adjustment of the model’s parameters so that the output will be comparable to the data obtained from field experiments. We first calibrated soil parameters, then the genetic coefficient for the cowpea variety and finally, the experimental data. CROPGRO-Cowpea model requires 15 cultivar coefficients (CSDL, PPSEN, EM-FL, FL-SH, FL-SD, SD-PM, FL-LF, LFMAX, SLAVR, SIZLF, XFRT, WTPSD, SFDUR, SDPDV, and PODUR) that describe the growth and development characteristics for each individual cultivar The cultivar coefficients for the three cultivars UAM091051-1, IT99K-573-1-1 and IT90K-277- 2 were generated from the existing cultivars in DSSAT. The existing cultivars with the same maturity group and characteristics of a tropical cowpea varieties were used to generate the coefficients. IT96D-748F was used for IT99K-573-1-1 while IT90K-277-2 already in the DSSAT was used as template for the IT90K-277-2 and UAM09 1051-1. The cultivar coefficients for each cultivar were determined through trial and error of the model and by comparing simulated and observed data, following the procedures described by Hoogenboom et al. (1999). 33 3.3.1.2. Model validation The performance of the model was validated with the four independent data sets collected from the population density trials for the three cowpea varieties using three levels (133,333, 266,666 and 400,000 Plants ha-1). The experiments were conducted from 2016 to 2018 at Abuja and Zaria in the northern Guinea savanna zone. The treatment consisted of population density and cowpea varieties. A split-plot design with three replications was used. The main plot treatment consisted of three population density (133,333, 266,666 and 400,000 Plants ha-1) levels and the subplot treatments was assigned to three cowpea (IT99K-573-1-1, IT90K-277-2 and UAM 09 1051-1) varieties. Soil and weather data were collected from the experimental sites and used as inputs for model validation. For validation of the model, only grain yield outputs of the model were compared with observed grain yields obtained from the validation experiments. An analysis of the degree of coincidence between simulated and observed values was statistically determined with the following methods: (i) Root Mean Square Error (RMSE), this reflects the extent of the mean variance between simulated and observed data and it is a good measure of how accurately the model predicts the response. A good value of the RMSE should approach zero (Halder et al. 2017). (ii) Wilmott index of agreement (d-index) Willmott (1982), the nearer the value of d to 1, the better the prediction. 3.3.1.3. Model application (Seasonal analysis) The model was applied to test the performance of different cowpea varieties in 18 and 15 communities in Adamawa and Borno States, respectively, using the seasonal analysis tool of DSSAT v4.7. The seasonal analysis was set to use July 20 as sowing date for all varieties at both States. A Single Super Phosphate (SSP) fertilizer was used in the model to supply 40 kg P2O5 at planting. The model was set to use 3 cm soil depth, with a recommended sowing density of 26.6 plants per square meter. The model was set to harvest when the crop reached harvest maturity. 30-year weather records (1985-2014) obtained from NIMET was used for seasonal analysis. Soil profile data for Adamawa and Borno were used for the scenario analysis. The mean grain yields for the 30 years for each variety and location were calculated using DSSAT. 3.3.2. Results 3.3.2.1 Model calibration The results for the model calibration experiments for cowpea varieties are presented in Figures 5- 7 A close agreement was obtained between simulated and observed values for all three measured parameters. The statistical values of simulated and measured days to anthesis and physiological maturity ranged between (RMSE = 0.82 to 2.37 days) and (d-index = 0.76 to 0.96) for all varieties. The comparison between simulated and observed grain yield were also quite good for all the cowpea varieties. The RMSE values for grain yield ranged between 123 kg ha-1 for UAM 09 1051-1 to 270 kg ha-1 for IT99K-573-1-1 while d-index value was 0.77, 0.81 and 0.89 for IT99K- 573-1-1, IT90K-277-2 and UAM 09 1051-1, respectively. 34 Fig. 3.5 Comparison of simulated and measured anthesis (a), physiological maturity (b) and grain yield at maturity (c) for IT99K-573-1-1 cowpea variety. y = 0.4356x + 23.861 R² = 0.599 y = x 0 5 10 15 20 25 30 35 40 45 50 0 10 20 30 40 50 A n th e si s d ay ( d ap ) m e as u re d Anthesis day (dap) simulated a RMSE =1.22 D =0.79 y = 0.4435x + 41.461 R² = 0.4435 y = x 0 10 20 30 40 50 60 70 80 90 0 20 40 60 80 P h ys io lo gi ca l m at u ri ty d ay ( d ap ) m e as u re d Physiological maturity day (dap) simulated b RMSE =2.31 D =0.78 y = 0.7791x + 698.8 R² = 0.4597 y = x 0 500 1000 1500 2000 2500 3000 3500 0 1000 2000 3000 Y ie ld a t h ar ve st m at u ri ty ( kg [ d m ]/ h a) m e as u re d Yield at harvest maturity (kg [dm]/ha) simulated c RMSE =269.76 D =0.77 35 Fig. 3.6 Comparison of simulated and measured anthesis (a), physiological maturity (b) and grain yield at maturity (c) for IT90K-277-2 cowpea variety. y = x + 0.3333 R² = 0.84 y = x 0 10 20 30 40 50 60 0 10 20 30 40 50 60 A n th e si s d ay ( d ap ) m e as u re d Anthesis day (dap) simulated a RMSE =0.82 D =0.95 y = 0.8655x + 10.345 R² = 0.9064 y = x 0 10 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 P h ys io lo gi ca l m at u ri ty d ay ( d ap ) m e as u re d Physiological maturity day (dap) simulated b RMSE =1.73 D =0.94 y = 1.3354x - 691.76 R² = 0.7984 y = x 0 500 1000 1500 2000 2500 3000 3500 0 1000 2000 3000 Y ie ld a t h ar ve st m at u ri ty ( kg [ d m ]/ h a) m e as u re d Yield at harvest maturity (kg [dm]/ha) simulated c RMSE =190 D =0.81 y = 0.4708x + 45.942 R² = 0.3893 1:1 line 0 10 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 P h ys io lo gi ca l m at u ri ty d ay ( d ap ) m e as u re d Physiological maturity day (dap) simulated b RMSE =3.37 D =0.76 y = 0.7143x + 15.524 R² = 0.6957 1:1 line 0 10 20 30 40 50 60 70 0 20 40 60 A n th e si s d ay ( d ap ) m e as u re d Anthesis day (dap) simulated a RMSE =1 D =0.9 36 Fig. 3.7 Comparison of simulated and measured anthesis (a), physiological maturity (b) and grain yield at maturity (c) for UAM 09 1051-1 cowpea variety. 3.3.2. 2 Model validation The accuracy of the CROPGRO model simulations and performance of genetic coefficients were assessed by running the model with independent data sets collected during 2016, 2017 and 2018 seasons for IT99K-573-1-1, IT90K-277-2 and UAM 09 1051-1using three population densities (133,333, 266,666 and 400,000) at Zaria. Grain yield at maturity was used for model evaluation (Table 7). The model's evaluation of grain yield was good at all population densities of the three varieties. In all the years, the model slightly underestimated or overestimated the measured parameters for all the three cowpea varieties at various population densities. However, the under or over estimations were within the acceptable range of below 25 %. There was a good fit in the model prediction of grain yield with low RMSE and high d-index values. The values of RMSE for the three varieties ranged from 118.1 to 176.3 kg ha-1. In all cases, d-index values for grain yield were above 0.80 indicating that the model is robust and accurate in measuring grain yield. Table 3.7 Simulated (S), observed (O) and simulated minus observed (S - O) of cowpea grain yield obtained from validation experiments conducted at Zaria over a three-year period. Location Year Density IT99K-573-1-1 IT90K-277-2 UAM 09 1051-1 S O S - O S O S - O S O S - O Zaria 2016 133,333 1860 1818 42 1522 1376 146 1575 1337 238 266,666 1864 1920 -56 1421 1413 8 1593 1363 230 400000 1879 1899 -20 1433 1400 33 1672 1925 -253 Zaria 2017 133,333 1483 1602 -119 1936 1718 218 1969 2021 -52 266,666 1504 1742 -238 2004 2185 -181 2080 2120 -40 400000 1537 1712 -175 2102 2081 21 2167 2074 93 Zaria 2018 133,333 1918 1885 33 2314 2207 107 1575 1337 238 266,666 1945 1878 67 2366 2400 -34 1593 1363 230 400000 1602 1717 -115 2381 2296 85 1672 1925 -253 RMSE 118.1 129.2 176.31 d value 0.83 0.96 0.9 d, Willmott index of agreement (Willmott, 1982) ranging from 0 to 1, 1 being perfect agreement. y = 1.228x - 465.58 R² = 0.7767 1:1 line 0 500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500 3000 Y ie ld a t h ar ve st m at u ri ty ( kg [ d m ]/ h a) m e as u re d Yield at harvest maturity (kg [dm]/ha) simulated c RMSE =123 D =0.89 37 3.3.2.3 Model application (Seasonal analysis) Results of seasonal analysis for mean grain yields in different locations conducted by the DSSAT model over 30-year period are presented in Tables 8 and 9 for Borno and Adamawa, respectively. In the two States, location had significant long-term effect on grain yield of all the varieties. On average, the varieties produced higher grain yield in Borno than in Borno with the medium maturing Striga resistant variety (IT99K-573-1-1) recording the highest grain yield of above 1- ton ha-1. Average yield across locations, Mathau in Biu gave the highest grain yield of 1116 kg ha-1 and the lowest grain of 960 kg ha-1 was obtained at Bila Gusi in Kwayakusar in Borno State while Guyaku at Gombi in Adamawa, gave the highest grain yield of 1101 kg ha-1 and the lowest grain yield of 731 kg ha-1 was observed at Yola north. Overall, the model simulated low yields for all cowpea varieties when compared with yields obtained under experimental conditions. We will continue to work with the developers of the model to refine it for better simulation of cowpea performance. Because of the low incidence of the parasitic weed Striga gesneroides in Adamawa State, all the three cowpea varieties can be recommended for production there. In Borno State where Striga is a problem, the two varieties (IT99K-573-1-1 and UAM 09 1051-1) which are completely resistant to Striga are recommended for dissemination. Table 3.8 Cowpea grain yield (kg ha-1) results of 30 years (1985-2014) seasonal analysis in Borno, northeast Nigeria. LGA Community IT99K-573-1-1 IT90K-277-2 UAM 09 1051- 1 Mean Bayo Balbaya 1102.5 1035.1 1010.5 1049.4 Bayo Briyel 1113.6 1037.3 1020.8 1057.2 Bayo Jara-Dali 1111.6 1036.6 1020.0 1056.1 Biu Buratai 1076.8 1010.4 991.3 1026.2 Biu Kabura 1108.6 1034.4 1037.2 1060.1 Biu Tum 1088.4 1017.8 976.4 1027.5 Biu Mathau 1115.9 1033.7 1040.3 1063.3 Hawul Kwajaffa 1092.4 1038.0 1037.7 1056.0 Hawul Puba Vidau 1088.3 1039.3 1038.4 1055.3 Hawul Sakwa Hema 1092.4 1038.0 1037.7 1056.0 Kwayakusar Bila Gusi 1004.8 962.3 960.2 975.8 Kwayakusar Kurbo Gayi 1113.4 1037.7 1021.6 1057.6 Shani Gwaskara 1108.1 1041.1 1044.2 1064.5 Shani Kubo 1007.9 962.2 962.2 977.4 Shani Lakundum 1107.5 1041 1044 1064.2 Mean 1088.8 1024.3 1016.2 38 Table 3.9 Cowpea grain yield (kg ha-1) results of 30 years (1985-2014) seasonal analysis in Adamawa, northeast Nigeria. LGA Community IT99K-573-1-1 IT90K-277-2 UAM 09 1051-1 Mean Demsa Mbula Kuli 1060.4 970.7 964.8 998.6 Demsa Nassarawo-Demsa 1022.7 973.3 961.9 986.0 Girei Daneyel 996.8 947.4 940.6 961.6 Girie Wuroshi 1009.5 938.1 933.2 960.3 Gombi Tawa 1085.2 1001.7 984.8 1023.9 Gombi Guyaku 1100.5 1067.3 1057.5 1075.1 Guyuk Chikila 992.0 952.5 951.9 965.5 Guyuk Lakumna 1010.5 961.3 962.0 977.9 Hong Dulmava 1039.5 1003.1 1003.4 1015.3 Hong Garari_Hushere Zum 1071.7 1000.2 983.9 1018.6 Numan Bare 1063.3 948.6 936.8 982.9 Numan Kikan_Kodomti 1012.8 968.8 956.0 979.2 Shelleng Jonkolo – Lama 1011.3 958.9 955.1 975.1 Shelleng Lakati_Libbo 1051.3 964.7 951.6 989.2 Song Sabon Gari 1070.9 979.7 975.5 1008.7 Song Suktu 1070.9 979.7 975.5 1008.7 Yola North Yelwa -Jambore 731.1 753.3 771.3 751.9 Yola South Fufore 1072.5 1018.7 1007.3 1032.8 Mean 1026.3 966 959.6 3.4. Calibration, evaluation and application of the CROPGRO-Soybean model in DSSAT to simulate performance of soybean varieties in selected communities in Adamawa and Borno States 3.4.1. Methodology 3.4.1.1. Model calibration For the calibration of the model for soybean, 8 experiments were conducted across two sites at Bayero University, Kano and ABCOA, Danbatta in the Sudan savannas of Nigeria. Prior to the establishment of the experiment, soil samples were collected from the sites and analysed for nutrient content. Each experiment consisted of five soybean (TGX1904-6F, TGX 1951-3F, TGX1835-10E, TGX1987-10F and TGX1448-2E) varieties that