THESE DE DOCTORAT Présentée et soutenue publiquement pour obtention du titre de DOCTEUR des Université Paris VI- Pierre et Marie Curie & Académie de la Science et de la Technologie du Vietnam (Co-tutelle) Spécialité: Biogéochimie des hydrosystèmes Ecole Doctorale : Géoscience et Ressources Naturelles LE Thi Phuong Quynh FONCTIONNEMENT BIOGEOCHIMIQUE DU FLEUVE ROUGE (NORD –VIETNAM) : BILANS ET MODELISATION Soutenue le 7 Juillet 2005 Composition du jury: M. Venu ITTEKKOT Prof. Dr, CTME, Bremen, Allemagne Rapporteur M. Quang Cu BUI Prof. Dr, VAST, HoChiMinh, Vietnam Rapporteur M. Georges VACHAUD Prof. Dr, CNRS, Grenoble, France Rapporteur M. Ghislain DE MARSILY Prof., Univ. Paris VI, Paris, France Examinateur M. Wolfgang LUDWIG Dr, CEFREM, Perpignan, France Examinateur Mme. Josette GARNIER Dr, CNRS- Univ. Paris VI, Paris, France Directrice de thèse M. Gilles BILLEN Dr, CNRS- Univ. Paris VI, Paris, France Directeur de thèse M. Van Minh CHAU Prof. Dr, VAST, Hanoi, Vietnam Co-Directeur de thèse Thèse préparée au sein des laboratoires Sisyphe, UMR 7619, CNRS (France) – INPC, VAST (Vietnam) THESIS Written and defended for obtaining the doctorate degree of Pierre et Marie Curie University (France) & Vietnamese Academy of Science and Technology (Vietnam) (Co-supervision) Speciality: Biogeochemistry of hydrosystems PhD School: Geoscience and Natural Resources LE Thi Phuong Quynh BIOGEOCHEMICAL FUNCTIONING OF THE RED RIVER (NORTH VIETNAM): BUDGETS AND MODELLING Defended on July 7th 2005 Composition of the Committee: Reporter M. Venu ITTEKKOT Prof. Dr, CTME, Bremen, Germany Reporter M. Quang Cu BUI Prof. Dr, VAST, HoChiMinh, Vietnam Reporter M. Georges VACHAUD Prof. Dr, CNRS, Grenoble, France Examinator M. Ghislain DE MARSILY Prof., Univ. Paris VI, Paris, France Examinator M. Wolfgang LUDWIG Dr, CEFREM, Perpignan, France Advisor Mme. Josette GARNIER Dr, CNRS- Univ. Paris VI, Paris, France Advisor M. Gilles BILLEN Dr, CNRS- Univ. Paris VI, Paris, France Co- Advisor M. Van Minh CHAU Prof. Dr, VAST, Hanoi, Vietnam This thesis is prepared at the laboratories Sisyphe, UMR 7619, CNRS (France) – INPC, VAST (Vietnam) Acknowledgements Acknowledgements First of all, I am extremely grateful to my advisors Dr. Josette Garnier and Dr. Gilles Billen for accepting me as their PhD student and for their enormous assistance and helpful discussions during my thesis. They have helped me in understanding concepts in a simple and intuitive way, while introducing me to new ideas. They always know how to solve the problems and always encourage me during the difficult periods. They have always offered me special cares during my stays in France so that I could feel happy and comfortable. I would express my particularly thanks to them. I would like to thanks my Vietnamese co-advisor, Prof. Dr. Chau Van Minh who gives me the opportunity to work in the ESPOIR project and to realize the cotutelle Ph.D. thesis. He always provides me the favorable working conditions in the Institute of Natural Products Chemistry (INPC). Without his helps in experiments, samplings and administrative papers in INPC in Vietnam, the thesis would never be finished. The PhD thesis was performed in the ESPOIR project, a French-Vietnamese program for water quality and water treatment in the period from 2000 to 2004. I would like to thank Prof. Georges Vachaud, Prof. Chau Van Minh and Prof. Nguyen The Dong to give me the chance to pursue this thesis in the framework of the ESPOIR project. I am also indebted to Prof. Ghislain de Marsily, the ex-director of the Ecole Doctorale “Géosciences et Ressources Naturelles” for analysing my Vietnamese degree courses and accepting my inscription. I must also thank Prof. Laurent Jolivet, the present director of the Ecole Doctorale for his kindness with the administrative forms that permits my continuing during the last period of this thesis. Furthermore, I am deeply thankful to all the members of the jury: Prof. Venu Ittekkot, Prof. Georges Vachaud, Prof. Bui Quang Cu, Prof. Ghislain De Marsily, Dr Wolfgang Ludwig, Dr. Josette Garnier, Dr Gilles Billen and Prof. Chau Van Minh, who gave many interesting and helpful comments and critics for my thesis manuscript and also for the enrichment of my scientific knowledge. During this work, I have been granted by the French Embassy in Vietnam at Hanoi. I would like to express my thanks to the French Embassy in Vietnam and I especially thank Mr Bruno Paing, attached to the cooperation of Science and Technology for his interest in this programme and for always helping me with kindness in finding administrative solutions. This work is a cotutelle thesis. I would also like to acknowledge the Leaders of Institute of Natural Products Chemistry, the Leaders of University of Pierre and Marie Curie, who i Acknowledgements permitted me to carry out this work. Helpful financial supports was provided by the Direction of the International Cooperation of the Pierre and Marie Curie University. I wish to extend a sincere gratitude to the director of Sisyphe laboratory, Prof. Alain Tabbagh, to give me the warm welcome in this laboratory. I express my sincere thanks to Sylvain Théry, a very humorous, friendly and hard working person, for his huge helps, especially in the Red River data base elaboration, logical programs creation and map drawing. Moreover, I would like to thank the kind colleagues Nguyen Van Tuan, Tran Bich Nga and Nguyen Van Tue in Meteorological and Hydrological Institute for their useful helps in Vietnamese meteo-hydrological information. I would like to thank the sympathetic colleagues in the Son Tay, Yen Bai, Hoa Binh, Vu Quang hydrological stations for their helps in water samplings. Among all the numerous people who have contributed to valuable ideas and experiments related to this work, I would like to mention Dr Michel Meybeck and Dr Agnes Ducharne (Sisyphe), Dr Pham Van Cu (Institute of Geography in Vietnam), Dr. Pham Huu Dien (Hanoi, Pedagogic University I), Dr Nguyen Kien Cuong, Prof. Ngo Ngoc Cat and Dr Nguyen Thanh Van (VAST). I express a deep gratitude to them. I also wish to express my gratitude to my colleagues in the Institute of Natural Products Chemistry: Luong, Thao, M. Ha; they have given so much help in the sampling campaigns and sample analyses. My thanks are also due to Nicolas Prieur, who spent two years as a CNRS Engineer making the link between French and Vietnamese team, and had a large contribution in the organisation of the sampling campaigns for Nhue-Tolich urban rivers in the framework of the ESPOIR project. At Sisyphe, I truly thank Nadine and Valérie at the management and secretaryship, Maya responsible for the informatics. I would like to sincerely acknowledge the generous assistance provided by the following colleagues: Maïa, Séverine, Anun, Mohamed, Samia, Maïté. My thanks are sent also to all these so kind friends: Agata, Véronique, Aurélie, Julien, Harouna, Denis, Anne, Angelbert, Hans, Noémi, D. Thuy, Tam … for cheering me during the four half- year stays in France. At last but above all, I would like to extend my sentiments to my closest relatives. I am greatly indebted to my parents and sisters for their morale and love supports as well as their confidence in my scientific orientations and decisions. I am happy to get diploma, but my parents are proud of that. I am deeply grateful to my husband who no only always understands, believes and encourages me, but also helps me. My best Vietnamese friends Binh, Trang, Long, Phong, Vu, Loi, Thuc are thanked for the wonderful days we spent together at school and/or University and for their continuous encouragements. ii Résumé Résumé Le Fleuve Rouge (au Nord Vietnam et en Chine méridionale) couvre une surface de bassin versant de 156 450 km2, avec une population de près de 30 millions d’habitants. L’axe principal du Fleuve Rouge (aussi appelé Yuan, Thao ou Hong) reçoit deux affluents principaux, le Da et le Lo, puis forme un large delta avant de se jeter dans le Golfe du Tonkin (en Mer de Chine méridionale). Les trois sous-bassins supérieurs et le delta diffèrent largement en terme de densité de population (de 101 hab.km-2 dans les bassins amont à plus de 1000 hab.km-2 dans le delta), d’usage du sol et de pratiques agricoles. Le but général du présent travail est de développer une compréhension d’ensemble du fonctionnement biogéochimique de ce système sub-tropical de dimension régionale, et de son contrôle par les processus naturels et anthropiques. L’épine dorsale du travail a consisté dans l’implémentation du modèle RIVERSTRAHLER, développé antérieurement pour décrire le lien entre la qualité de l’eau et les activités humaines dans le bassin de la Seine et d’autres fleuves européens (Billen et al., 1994, 1997, 1999, 2005; Garnier et al., 1995, 1999, 2000, 2002), pour le cas particulier du système Fleuve Rouge. La première étape dans cette étude a consisté dans la modélisation du régime hydrologique et du transport solide du Fleuve Rouge (Le Thi Phuong Quynh et al., subm). Les estimations antérieures de la charge solide du Fleuve Rouge variaient entre 100 et 170 106 t.an-1, c-à-d de 640 à 1060 t.km-².an-1. La forte dépendance du transport solide à l’hydrologie est responsable d’une large variabilité inter-annuelle. Sur la base de données hydrologiques relatives à la période 1997-2004, et d’un suivi journalier de la matière en suspension à l’exutoire des 3 principaux tributaires du Fleuve Rouge en 2003, un modèle simplifié a été établi pour estimer la charge solide moyenne interannuelle du Fleuve Rouge sous les conditions actuelles. La valeur obtenue est de 40 106 t.an-1, correspondant à une charge spécifique de 280 t.km-2.an-1. Elle reflète une réduction de 70% de la charge solide totale suite à la mise en eau des réservoirs de Hoa Binh et de Thac Ba réservoirs dans les années 1980s. Le modèle prévoit une réduction supplémentaire de 20% de la charge en suspension suite à la construction planifiée de deux grands réservoirs supplémentaires. Utilisant les mesures de contenu en phosphore total dans la matière en suspension réalisées dans ce travail, le flux de phosphore exporté par le Fleuve Rouge peut être estimé à 36 106 kgP an-1. Les données de concentrations en nutriments dans le réseau hydrographique du Fleuve Rouge étant assez rares, un suivi de la concentration des formes de l’azote, du phosphore, de la silice, du carbone organique et de la chlorophylle à l’exutoire des principaux sous-bassins amont, dans l’axe principal du Fleuve dans le delta et dans quelques rivières polluées de la iii Résumé région d’Hanoï, a été réalisé à une fréquence mensuelle durant les années 2003 et 2004, permettant de définir le niveau général de concentration en nutriments dans les eaux de surface. En vue d’examiner le degré de perturbation anthropique du cycle de l’azote et du phosphore à l’échelle du bassin, des bilans de ces deux éléments ont été établis pour le système sol et pour l’hydrosystème des 4 principaux sous-ensembles (Da, Lo, Thao et Delta) du bassin du Fleuve Rouge (Le Thi Phuong Quynh et al., 2005). En terme de production agricole, d’une part, de consommation de nourriture et de fourrage d’autre part, les sous-basins amont apparaissent comme des systèmes autotrophes, exportant des produits agricoles, tandis que le delta dépend d’importations de biens agricoles. Le bilan des sols agricoles révèle de fortes pertes d’azote, principalement attribuables à la dénitrification dans les rizières, et de phosphore, principalement dues à l’érosion. Le bilan du réseau hydrographique montre une importante rétention/élimination d’azote (de 62 à 77 % dans les basins amont et de 59 % dans le delta), et de phosphore, avec un taux de rétention de plus de 80 % dans le Da et le Lo, à l’aval desquels sont localisés les grands réservoirs (Hoa Binh sur le Da et Thac Ba sur le Lo). L’exportation spécifique estimée à l’exutoire du Fleuve Rouge est estimée à 855 kg.km-².an-1 d’azote total et 325 kg.km-².an-1 de phosphore total. L’azote plutôt que le phosphore semble être l’élément limitant principal de la croissance algale dans les zones côtières influencées par le Fleuve Rouge dans le Golfe du Tonkin. Une base de données sous SIG a été assemblée à l’échelle du bassin du Fleuve Rouge, avec des couches d’informations renseignant la géomorphologie du bassin, sa lithologie, la météorologie, l’usage du sol et les pratiques agricoles, la population et les rejets d’eau usées domestiques et industrielles. Cette base de données est conforme au format requis par le logiciel SENEQUE/Riverstrahler (Ruelland et al, 2004), une version du modèle Riverstrahler encapsulée dans une interface SIG constituant un outil de modélisation générique et spatialement explicite de la qualité de l’eau à l’échelle des grands réseaux hydrographiques. La première application de ce logiciel au système Fleuve Rouge est décrite et validée sur la base des données acquises lors des suivis mensuels de qualité d’eau à l’exutoire des grands sous-bassins et sur l’axe principal du Fleuve lors des années 2003 et 2004. Enfin, le modèle a été utilisé pour explorer l’effet, en terme de qualité de l’eau et de fonctionnement biogéochimique de divers scénarios décrivant de possibles changements futurs du bassin du Fleuve Rouge concernant son aménagement hydraulique, l’usage de ses sols et son agriculture, sa population et sa gestion des eaux usées. Mots clés: Rivière Tropicale, Fleuve Rouge, Vietnam, modèle Riverstrahler/Sénèque, nutriments, cycle de l’azote, du phosphore, de la silice, charge solide. iv Résumé Summary The Red River (in North Vietnam and South China) covers a watershed area of 156 450 km2 with a total population near 30 million inhabitants. The main branch of the Red River (also called Yuan, Thao or Hong River) receives two major tributaries, the Da and Lo Rivers, then forms a large delta before discharging into the Tonkin Bay (South China Sea). The 3 upstream sub-basins and the Delta area differ widely in population density (from 101 inhab km-2 in the upstream basins to more than 1000 inhab km-2 in the delta), land use and agricultural practices. The general goal of this work is to develop a comprehensive understanding of the biogeochemical functioning of this sub-tropical regional system, and its control by natural and anthropogenic processes. The backbone of the work consisted in implementing the RIVERSTRAHLER Model, previously developed for describing the link between water quality and human activities in the watershed in the Seine river and other European river systems (Billen et al., 1994, 1997, 1999, 2005 ; Garnier et al, 1995, 1999,2000, 2002) to the special case of the Red River system. The first step of the study consisted in modeling the hydrological regime and the suspended solid transport of the Red River (Le Thi Phuong Quynh et al., subm). Previous estimates of its suspended matter loading range from 100 to 170 106 t.yr-1, i.e. from 640 to 1060 t.km-².yr-1. The strong dependence of suspended solid transport on hydrology results in a large year-to- year variability. Based on available data on the hydrology over the period 1997-2004, and on one -year survey of the daily suspended matter of the three main tributaries of the Red River system in 2003, a simplified modeling approach is established to estimate the mean suspended loading of the Red River under present conditions. The obtained value is 40 106 t.yr-1, corresponding to a specific load of 280 t.km-2.yr-1. It reflects a 70% decrease of the total suspended load since the impoundment of the Hoa Binh and Thac Ba reservoirs in the 1980’ies. The model predicts a further reduction by 20% of the suspended loading of the Red River with the planned construction of two additional reservoirs. Using measurements of the total phosphorus content of the suspended material in the different Red River tributaries, we could estimate the present phosphorus delivery by the Red River as 36 106 kgP yr-1. As data on nutrient concentration in the Red River drainage network are rather scarce, a survey of nutrient concentration (N, P, Si, organic carbon and chlorophyll a) at the outlet of the three main sub-basins, the main branch in the delta and some polluted rivers in the Hanoi v Résumé region was carried on at monthly intervals in 2003 and 2004, allowing to define the general levels of nutrient concentrations in surface water. In order to examine the degree of human-induced alteration of the nitrogen and phosphorus cycles at the scale of the watershed, budgets of these elements were established for the soil and the drainage network of the 4 main sub-basins (Da, Lo, Thao and Delta) of the Red River (Le Thi Phuong Quynh et al., 2005). In terms of agricultural production, on the one hand, and consumption of food and feed on the other, the upstream sub-basins are autotrophic systems, exporting agricultural goods, while the delta is a heterotrophic system, depending on agricultural goods imports. The budget of the agricultural soils reveals great losses of nitrogen, mostly attributable to denitrification in rice paddy fields and of phosphorus, mostly caused by erosion. The budget of the drainage network shows high retention/elimination of nitrogen (from 62 to 77 % in the upstream basins and 59 % in the delta), and of phosphorus, with retention rates as high as 80 % in the Da and Lo sub-basins which have large reservoirs in their downstream course (Hoa Binh on the Da and Thac Ba on the Lo). The total specific delivery estimated at the outlet of the whole Red River System is 855 kg.km-².y-1 total N and 325 kg.km-².yr-1 total P. Nitrogen rather than phosphorus seems to be the potential limiting factor of algal growth in the plume of the Red River in Tonkin Bay. A GIS data base has been assembled at the scale of the whole Red River basin, with layers documenting geomorphology, lithology, meteorology, land-use and agriculture, population, domestic and industrial wastewater release, etc. This data base follows the format required for running the SENEQUE/Riverstrahler software (Ruelland et al, 2004), a version of the Riverstrahler model encapsulated into a GIS interface in order to build a generic and spatially explicit water quality modelling tool. The first application of this model to the Red River system is described and validated with the data acquired by the monthly surveys of water quality at the outlet of the 3 sub-basins and in the main branch of the Red River during the years 2003 and 2004. Finally, the model is used to explore the effect in terms of water quality and biogeochemical functioning of a variety of scenarios describing possible future changes in the Red River basin concerning hydrological management, land use and agricultural practices, population increase and wastewater treatment policy. Key words: tropical river, Red River, Vietnam, Riverstrahler/Seneque model, nutrient budgets, nitrogen, phosphorus, silica cycle, suspended solids. vi