The relationship between vegetation associations and floods Vegetation associations have privileged relationships with   flood heights, as shown by Pierre Hiernaux , in conclusion of the statistical analyses establishing the floristic profiles of these associations and the statistical relationships between ecological variables and vegetation associations: the flood parameters appear as the most determinant variables (Hiernaux et al . 1980). The privileged role of the flood is confirmed by the overall scheme of ecological relationships, which results from a factorial analysis of the matrix of the state of environment variables and the presence-absence of vegetation species. Indeed, whether the analysis focuses on all 169 survey sites or on a selection of the 127 sites within the proper flood plains, the structure of "matrix clouds" is characterized by the strong hierarchy of the first axes, the main beams of the architecture of matrix clouds which constitute a synthetic representation of the interrelations – in terms of presence/absence – between taxa and states of the variable. In addition to the participation (expressed in %) of the first five axes in the inertia of the matrix, the following table shows the contribution of each of the variables to this participation, axis by axis, and overall for the five axes. Table 1: Features of the factor analysis of species / status variables for 169 readings of the Inner Niger Delta In the analysis of the 169 survey sites, the first two axes are closely linked to the flood conditions, which total 71.6% and 73.1% of the inertia explained by the first two axes. However, a more detailed observation shows that on the first axis, 46.3% of the inertia is explained by the ‘no flooding’ situation. In other words, the first axis opposes the non-floodable sites of the Delta to all of the flood plain sites, the second axis distributes the latter over a gradient of height, regularity and duration of flooding. (P. Hiernaux, The fodder resources map of the routes of the Inner Niger Delta - Notice , Bamako: CIPEA-ODEM, 1980, p.23) One of the main factors in the flood-vegetation relationship is the flood height, which makes it possible to classify the different vegetation associations by level, each level representing the most frequent flood height for a plant association. Table 9 : Plant associations according to the levels or heights of submersion At level 7, BP and PAK are the two plant formations that support – or require – the strongest submersions. The first is a low " bourgoutière" with Vossia cuspidata , the second corresponds to a very deep grove with Acacia kirkii . Level 6 is occupied by B a " bourgoutière" with Echinochloa stagnina , VB a very low " vetiveraie ", OP a deep "orizaie" and PAM a formation with Mitragina inernis . Level 5 is occupied by an "eragrostaie" EOR, an "orizaie" O and a "vétivéraie" VOR as well as rice paddies from the Office du Niger, downstream from Ké Macina. Level 4 is occupied by VSP and ESP, the "vetiveraies" and medium-sized "eragrostaies" as well as AC an "eragrostaie" with Andropogon canaliculatus . Level 3, which corresponds to a submersion between 10 cm and 30 cm, is occupied by a "panicaie" P and a high vétiveraie VH. Level 2 – the last level of submersion – between 0 and 10 cm, is occupied by AG, a savanna with Andropogon gayanus and a complex formation called ZB (for zone beaten by maximum floods). Finally –level 1 – the formations ranging from TA to TT are all located on the " togge ", the Fulani name for exposed mounds in the Delta and on the dry edges. Nb - PAN, PAS, PAR represent plant formations located on plains where flood is deferred. The most remarkable characteristic of these acacia-based formations is the bimodal dimension of the flood: the first mode is linked to the run-off of the rains from July- August, and the second to the late arrival of the flood in October or November. Finally, MB, the riverbank mosaic, is not shown in the table. It constitutes an alternation of channels and rims of banks or levees, well-represented in the major bed of the Niger, the Bani or the large tributaries. It always represents a strong gradient of submersion going from level 6 to level 2, juxtaposed over short distances (a few tens of meters). The definition of a "maximum reference flood" Pierre Hiernaux relates the different states of the variables describing the conditions of the flood (its height, regularity, duration, speed and the dates of the rise and fall of water…) with the data available in the Delta in 1980. The question is not to determine the regimes of each plain or of each basin taken in isolation, but to define a series of parameters which are most frequently reached, so as to make it possible to establish a relationship between plant formations and the course of a flood, which, by analogy to the relation existing between the climate and plant formations, can be considered as the "maximum reference flood". Pierre Hiernaux calculated the submersion depths normally attached to the flooded formations. These submergence depths are measured from a 0 cm point of reference, which corresponds to the level most regularly reached on the gauge. The comparison of the series leads us to consider that the value which can be considered as the reference value corresponds to the mean maximum of floods minus the standard deviation. P.A. Gosseye in (S. Cissé and P.A. Gosseye, Competition for limited resources: the case of the fifth region of Mali. Report 1: Natural resources and population. CABO, Wageningen, Netherlands - ESPR, Mopti, Mali. 1990, 106 p + appendices) discusses at length (pp. 48 and 49) this notion of a reference flood. In agreement with Pierre Hiernaux and with our own previous work (J. Marie, Search for a solution to the problems of breeding in the interior delta of Niger in Mali, vol.5, summary report., Addis Ababa , CIPEA / ODEM, 1983, 151 p. (1983). He admits, that for the Mopti station, the series of floods that can be described as normal and regular corresponds to the years 1944-1968, and that the average ten- day maximum for this period (686 cm), minus the standard deviation (26 cm), is the value most frequently reached or exceeded. This value therefore corresponds to the 660 cm level on the Mopti gauge (reached or exceeded in 84% of cases) and establishes, for the station of Mopti, a relation with the 0 cm reference for submergence heights. My own calculations relating to daily values ​​(and no longer ten-day ones) give very similar results: average 1943-1968: 693 cm; standard deviation 28 cm, i.e. a reference value of 665 cm reached in 82% of the cases. We will ultimately retain the value of 660 cm for the reference station of Mopti, a value which we assume to be representative of normal or regular floods, and which we therefore assimilate to the reference 0 cm of the height of submersion of the various flooded plant formations, which corresponds to the altitude of 267.20 m. This notion of "maximum reference flood", which suggests an equilibrium relationship between floods, varying from year to year, and plant formations can be discussed. Significant changes in floods should translate into changes in plant formations. The field work continued until 1985 (with a series of very low floods: 551 cm in 1982, 502 cm in 1983, 440 cm in 1984) shows very large variations in forage production, and limited modifications in the floristic composition of certain formations, but does not call into question the staggering of the vegetation composed of perennial grasses and the general pattern of the levels of submersion and their relationship with a reference flood. The conclusions of the return to the field of Pierre Hiernaux and Mathew Turner in 2014 leading to new vegetation surveys confirm this very great overall stability in the distribution of plant formations.(see page 43) Table 3 : Relationship between plant associations, submergence heights and flood at the Mopti gauge