Evaluating rice performance in contrasting East African wetlands using an experimental and modelling approach
Evaluating rice performance in contrasting East African wetlands using an experimental and modelling approach
View/Type of Scholarly Publication
DissertationDate of Exam
24.09.2021Date of Publication
07.10.2021Advisor
Becker, MathiasCo-Referee
Gaiser, ThomasMetadata
Show full item record
Citable Links 
Abstract
Rice (Oryza spp.) has become an important staple food and cash crop across East Africa, and is now playing a pivotal role in regional food security and national economies. Since regional rice yields remain low, intensification efforts increasingly focus on wetland areas due to favourable climatic and hydro-edaphic conditions. Among these focal areas are the Kilombero floodplain in southeast Tanzania and the numerous inland valley swamps in central Uganda.
We combined experimental and modelling approaches to assess differential yield responses to improved management and hydro-edaphic field conditions (hydrological floodplains' fringe and middle, and toposequence cross-section of the inland valley), and the use efficiency and profitability of N fertiliser using 3-years of experimental and 30-years of simulated data.
Against average regional lowland rice yields of <2 Mg ha-1, yields of between 3.2-9.2 Mg ha-1 in the floodplain and of between 1.9-6.3 Mg ha-1 in the inland valley were attained, depending on management (rainfed 0 and 60 kg N ha-1, irrigated 120 kg N ha-1 + 60 kg PK ha-1), field position and year. The APSIM model performed well for grain yield in both wetland types, as supported by numerous goodness-of-fit measures during model validation. Additionally, so-called 'carry-over' effects were simulated accurately, i.e., continuous soil moisture contents, soil carbon dynamics, and indigenous N supply. Results further emphasised the importance of water tables for lowland rice production and sound model performance as they evidently attenuated seasonal droughts. Furthermore, yield determinants from hydro-edaphic field conditions were delineated via simulated spatial-temporal abiotic stress patterns (water and N stress). In both wetlands, low soil N was the main constraint, while drought stress affected rice performance more in the inland valley, in years of low seasonal rainfall and from lower water table supply capacities. Long-term (30-years) scenario analysis of yield responses to and partial gross margins of N fertiliser indicated high fertiliser use efficiencies, i.e., with yield gains of 1.7-4.5 Mg ha-1 in the floodplain and 1.0-3.2 Mg ha-1 in the inland valley, and profitable N rates of 30-120 kg N ha-1 in the fringe and 30-90 kg N ha-1 in the middle position of the floodplain, and of 60-150 kg N ha-1 in the inland valley. However, fertiliser use was comparably riskier in the valley-fringe position of the inland valley. Supplemental irrigation effectively increased yields, particularly in the valley-fringe position and may help boost fertiliser use efficiency and profitability.
The APSIM model proved a valuable tool in prioritizing production constraints, assessing management options and thus in guiding the decision-making on crop management in wetlands. These finding may help to align regional self-sufficiency and conservation targets through site- and system-specific targeting of management options. Reis (Oryza spp.) hat sich in ganz Ostafrika zu einem wichtigen Grundnahrungsmittel und Cash-Crop entwickelt und spielt heute eine zentrale Rolle für die regionale Ernährungssicherheit und Volkswirtschaften. Da Reiserträge nach wie vor gering sind, konzentrieren sich Intensivierungsbemühungen aufgrund der günstigen klimatischen und hydro-edaphischen Bedingungen zunehmend auf Feuchtgebiete, u.a. die Kilombero-Überflutungsebene in Südost Tansania und die zahlreichen Talgrund-Feuchtgebiete in Zentraluganda.
Wir kombinierten experimentelle und Modellierungsansätze, um Ertragsreaktionen von Tieflandreis auf verbesserte Bewirtschaftungspraktiken und hydro-edaphische Feldbedingungen (hydrologischer Rand- und Mittebereich der Überflutungsebene, toposequentieller Querschnitt des Talgrund-Feuchtgebiets), sowie die Nutzungseffizienz und Rentabilität von Stickstoff (N)-Dünger anhand von 3-Jahren experimenteller und 30-Jahren simulierter Daten zu bewerten.
Gegenüber regionalen regenwassergespeisten Tieflandreiserträgen von <2 Mg ha-1 wurden Erträge zwischen 3,2-9,2 Mg ha-1 in der Überflutungsebene, und zwischen 1,9-6,3 Mg ha-1 im Talgrund-Feuchtgebiets erreicht, abhängig von Bewirtschaftung (regenwassergespeist 0 und 60 kg N ha-1, bewässert 120 kg N ha-1 + 60 kg PK ha-1), Feldposition und Jahr. Das APSIM-Modell zeigte eine gute Modellierleistung in beiden Feuchtgebietstypen, bestätigt durch zahlreiche Goodness-of-Fit-Güten während der Modellvalidierung. Zudem wurden "carry-over"-Effekte akkurat simuliert, d.h. Bodenfeuchtemuster, Bodenkohlenstoffdynamiken und die bodeneigene N-Versorgung. Ergebnisse betonten weiterhin die Bedeutung von Grundwasser für regenwassergespeisten Tieflandreis und eine gute Modellleistung. Simulierte räumlich-zeitliche abiotische Stressmuster (Wasser- und N-Stress) zeigten, dass geringe Bodenstickstoffgehalte die Hauptertragsbeschränkung war, während Trockenstress im Talgrund-Feuchtgebiet stärker ausgeprägt war. Die Szenario Analyse (30-Jahre) u.a. von partiellen Bruttomargen für N-Dünger ergaben profitable N-Raten von 30-120 kg N ha-1 am Rand und 30-90 kg N ha-1 in der Mitte der Überflutungsebene, und von 60-150 kg N ha-1 im Talgrund-Feuchtgebiet. Der Einsatz von N-Dünger war jedoch am Talrand vergleichsweise risikoreicher. Eine zusätzliche Bewässerung könnte dazu beitragen, die Effizienz und Rentabilität von N-Dünger zu steigern.
Das APSIM-Modell erwies sich als wertvolles Hilfsmittel bei der Priorisierung von Ertragsbeschränkungen, der Bewertung von Bewirtschaftungspraktiken und somit als Entscheidungshilfe für die landwirtschaftliche Nutzung von Feuchtgebieten. Ergebnisse können daher dazu beitragen, die regionale Selbstversorgung mit Reis und Ziele des Naturschutzes durch standort- und systemspezifische Bewirtschaftungspraktiken zu vereinbaren.
We combined experimental and modelling approaches to assess differential yield responses to improved management and hydro-edaphic field conditions (hydrological floodplains' fringe and middle, and toposequence cross-section of the inland valley), and the use efficiency and profitability of N fertiliser using 3-years of experimental and 30-years of simulated data.
Against average regional lowland rice yields of <2 Mg ha-1, yields of between 3.2-9.2 Mg ha-1 in the floodplain and of between 1.9-6.3 Mg ha-1 in the inland valley were attained, depending on management (rainfed 0 and 60 kg N ha-1, irrigated 120 kg N ha-1 + 60 kg PK ha-1), field position and year. The APSIM model performed well for grain yield in both wetland types, as supported by numerous goodness-of-fit measures during model validation. Additionally, so-called 'carry-over' effects were simulated accurately, i.e., continuous soil moisture contents, soil carbon dynamics, and indigenous N supply. Results further emphasised the importance of water tables for lowland rice production and sound model performance as they evidently attenuated seasonal droughts. Furthermore, yield determinants from hydro-edaphic field conditions were delineated via simulated spatial-temporal abiotic stress patterns (water and N stress). In both wetlands, low soil N was the main constraint, while drought stress affected rice performance more in the inland valley, in years of low seasonal rainfall and from lower water table supply capacities. Long-term (30-years) scenario analysis of yield responses to and partial gross margins of N fertiliser indicated high fertiliser use efficiencies, i.e., with yield gains of 1.7-4.5 Mg ha-1 in the floodplain and 1.0-3.2 Mg ha-1 in the inland valley, and profitable N rates of 30-120 kg N ha-1 in the fringe and 30-90 kg N ha-1 in the middle position of the floodplain, and of 60-150 kg N ha-1 in the inland valley. However, fertiliser use was comparably riskier in the valley-fringe position of the inland valley. Supplemental irrigation effectively increased yields, particularly in the valley-fringe position and may help boost fertiliser use efficiency and profitability.
The APSIM model proved a valuable tool in prioritizing production constraints, assessing management options and thus in guiding the decision-making on crop management in wetlands. These finding may help to align regional self-sufficiency and conservation targets through site- and system-specific targeting of management options.
Wir kombinierten experimentelle und Modellierungsansätze, um Ertragsreaktionen von Tieflandreis auf verbesserte Bewirtschaftungspraktiken und hydro-edaphische Feldbedingungen (hydrologischer Rand- und Mittebereich der Überflutungsebene, toposequentieller Querschnitt des Talgrund-Feuchtgebiets), sowie die Nutzungseffizienz und Rentabilität von Stickstoff (N)-Dünger anhand von 3-Jahren experimenteller und 30-Jahren simulierter Daten zu bewerten.
Gegenüber regionalen regenwassergespeisten Tieflandreiserträgen von <2 Mg ha-1 wurden Erträge zwischen 3,2-9,2 Mg ha-1 in der Überflutungsebene, und zwischen 1,9-6,3 Mg ha-1 im Talgrund-Feuchtgebiets erreicht, abhängig von Bewirtschaftung (regenwassergespeist 0 und 60 kg N ha-1, bewässert 120 kg N ha-1 + 60 kg PK ha-1), Feldposition und Jahr. Das APSIM-Modell zeigte eine gute Modellierleistung in beiden Feuchtgebietstypen, bestätigt durch zahlreiche Goodness-of-Fit-Güten während der Modellvalidierung. Zudem wurden "carry-over"-Effekte akkurat simuliert, d.h. Bodenfeuchtemuster, Bodenkohlenstoffdynamiken und die bodeneigene N-Versorgung. Ergebnisse betonten weiterhin die Bedeutung von Grundwasser für regenwassergespeisten Tieflandreis und eine gute Modellleistung. Simulierte räumlich-zeitliche abiotische Stressmuster (Wasser- und N-Stress) zeigten, dass geringe Bodenstickstoffgehalte die Hauptertragsbeschränkung war, während Trockenstress im Talgrund-Feuchtgebiet stärker ausgeprägt war. Die Szenario Analyse (30-Jahre) u.a. von partiellen Bruttomargen für N-Dünger ergaben profitable N-Raten von 30-120 kg N ha-1 am Rand und 30-90 kg N ha-1 in der Mitte der Überflutungsebene, und von 60-150 kg N ha-1 im Talgrund-Feuchtgebiet. Der Einsatz von N-Dünger war jedoch am Talrand vergleichsweise risikoreicher. Eine zusätzliche Bewässerung könnte dazu beitragen, die Effizienz und Rentabilität von N-Dünger zu steigern.
Das APSIM-Modell erwies sich als wertvolles Hilfsmittel bei der Priorisierung von Ertragsbeschränkungen, der Bewertung von Bewirtschaftungspraktiken und somit als Entscheidungshilfe für die landwirtschaftliche Nutzung von Feuchtgebieten. Ergebnisse können daher dazu beitragen, die regionale Selbstversorgung mit Reis und Ziele des Naturschutzes durch standort- und systemspezifische Bewirtschaftungspraktiken zu vereinbaren.
Subjects
APSIM, Oryza spp., Talgrund-Feuchtgebiet, Tansania, Uganda, Überflutungsebene, Alluvial floodplain, inland valley swamp, Tanzania
Classification (DDC)
630 Landwirtschaft, VeterinärmedizinGrotelüschen, Kristina: Evaluating rice performance in contrasting East African wetlands using an experimental and modelling approach. - Bonn, 2021. - Dissertation, Rheinische Friedrich-Wilhelms-Universität Bonn.
Online-Ausgabe in bonndoc: https://nbn-resolving.org/urn:nbn:de:hbz:5-63991
Online-Ausgabe in bonndoc: https://nbn-resolving.org/urn:nbn:de:hbz:5-63991
@phdthesis{handle:20.500.11811/9346,
urn: https://nbn-resolving.org/urn:nbn:de:hbz:5-63991,
author = {{Kristina Grotelüschen}},
title = {Evaluating rice performance in contrasting East African wetlands using an experimental and modelling approach},
school = {Rheinische Friedrich-Wilhelms-Universität Bonn},
year = 2021,
month = oct,
note = {Rice (Oryza spp.) has become an important staple food and cash crop across East Africa, and is now playing a pivotal role in regional food security and national economies. Since regional rice yields remain low, intensification efforts increasingly focus on wetland areas due to favourable climatic and hydro-edaphic conditions. Among these focal areas are the Kilombero floodplain in southeast Tanzania and the numerous inland valley swamps in central Uganda.
We combined experimental and modelling approaches to assess differential yield responses to improved management and hydro-edaphic field conditions (hydrological floodplains' fringe and middle, and toposequence cross-section of the inland valley), and the use efficiency and profitability of N fertiliser using 3-years of experimental and 30-years of simulated data.
Against average regional lowland rice yields of <2 Mg ha-1, yields of between 3.2-9.2 Mg ha-1 in the floodplain and of between 1.9-6.3 Mg ha-1 in the inland valley were attained, depending on management (rainfed 0 and 60 kg N ha-1, irrigated 120 kg N ha-1 + 60 kg PK ha-1), field position and year. The APSIM model performed well for grain yield in both wetland types, as supported by numerous goodness-of-fit measures during model validation. Additionally, so-called 'carry-over' effects were simulated accurately, i.e., continuous soil moisture contents, soil carbon dynamics, and indigenous N supply. Results further emphasised the importance of water tables for lowland rice production and sound model performance as they evidently attenuated seasonal droughts. Furthermore, yield determinants from hydro-edaphic field conditions were delineated via simulated spatial-temporal abiotic stress patterns (water and N stress). In both wetlands, low soil N was the main constraint, while drought stress affected rice performance more in the inland valley, in years of low seasonal rainfall and from lower water table supply capacities. Long-term (30-years) scenario analysis of yield responses to and partial gross margins of N fertiliser indicated high fertiliser use efficiencies, i.e., with yield gains of 1.7-4.5 Mg ha-1 in the floodplain and 1.0-3.2 Mg ha-1 in the inland valley, and profitable N rates of 30-120 kg N ha-1 in the fringe and 30-90 kg N ha-1 in the middle position of the floodplain, and of 60-150 kg N ha-1 in the inland valley. However, fertiliser use was comparably riskier in the valley-fringe position of the inland valley. Supplemental irrigation effectively increased yields, particularly in the valley-fringe position and may help boost fertiliser use efficiency and profitability.
The APSIM model proved a valuable tool in prioritizing production constraints, assessing management options and thus in guiding the decision-making on crop management in wetlands. These finding may help to align regional self-sufficiency and conservation targets through site- and system-specific targeting of management options.},
url = {https://hdl.handle.net/20.500.11811/9346}
}
urn: https://nbn-resolving.org/urn:nbn:de:hbz:5-63991,
author = {{Kristina Grotelüschen}},
title = {Evaluating rice performance in contrasting East African wetlands using an experimental and modelling approach},
school = {Rheinische Friedrich-Wilhelms-Universität Bonn},
year = 2021,
month = oct,
note = {Rice (Oryza spp.) has become an important staple food and cash crop across East Africa, and is now playing a pivotal role in regional food security and national economies. Since regional rice yields remain low, intensification efforts increasingly focus on wetland areas due to favourable climatic and hydro-edaphic conditions. Among these focal areas are the Kilombero floodplain in southeast Tanzania and the numerous inland valley swamps in central Uganda.
We combined experimental and modelling approaches to assess differential yield responses to improved management and hydro-edaphic field conditions (hydrological floodplains' fringe and middle, and toposequence cross-section of the inland valley), and the use efficiency and profitability of N fertiliser using 3-years of experimental and 30-years of simulated data.
Against average regional lowland rice yields of <2 Mg ha-1, yields of between 3.2-9.2 Mg ha-1 in the floodplain and of between 1.9-6.3 Mg ha-1 in the inland valley were attained, depending on management (rainfed 0 and 60 kg N ha-1, irrigated 120 kg N ha-1 + 60 kg PK ha-1), field position and year. The APSIM model performed well for grain yield in both wetland types, as supported by numerous goodness-of-fit measures during model validation. Additionally, so-called 'carry-over' effects were simulated accurately, i.e., continuous soil moisture contents, soil carbon dynamics, and indigenous N supply. Results further emphasised the importance of water tables for lowland rice production and sound model performance as they evidently attenuated seasonal droughts. Furthermore, yield determinants from hydro-edaphic field conditions were delineated via simulated spatial-temporal abiotic stress patterns (water and N stress). In both wetlands, low soil N was the main constraint, while drought stress affected rice performance more in the inland valley, in years of low seasonal rainfall and from lower water table supply capacities. Long-term (30-years) scenario analysis of yield responses to and partial gross margins of N fertiliser indicated high fertiliser use efficiencies, i.e., with yield gains of 1.7-4.5 Mg ha-1 in the floodplain and 1.0-3.2 Mg ha-1 in the inland valley, and profitable N rates of 30-120 kg N ha-1 in the fringe and 30-90 kg N ha-1 in the middle position of the floodplain, and of 60-150 kg N ha-1 in the inland valley. However, fertiliser use was comparably riskier in the valley-fringe position of the inland valley. Supplemental irrigation effectively increased yields, particularly in the valley-fringe position and may help boost fertiliser use efficiency and profitability.
The APSIM model proved a valuable tool in prioritizing production constraints, assessing management options and thus in guiding the decision-making on crop management in wetlands. These finding may help to align regional self-sufficiency and conservation targets through site- and system-specific targeting of management options.},
url = {https://hdl.handle.net/20.500.11811/9346}
}
The following license files are associated with this item:
