The   model   of   the   area’s   natural   environment   is   based   on   the   Niger   Delta   rangeland   map   produced   by   the   CIPEA-IER   /   ODEM team   as   explained   in   the   history   of   the   project.   It   covers   the   plains   of   the   Niger   basin   from   Ké   Macina   (on   the   Niger   River)   and Baramandougou   (on   the   Bani   River)   upstream   to   lake   Débo   downstream,   including   a   fraction   of   Farimaké   to   the   northwest   of   Lake Débo,   a   total   area   covering   slightly   over   22,000   km².   This   map,   together   with   the   accompanying   studies,   was   to   be   used   for   a regional   development   plan   with   the   creation   of   pastoralists'   or   agro-pastoralists'   associations.   This   perspective   determined   the   level of   precision   sought   in   the   definition   of   map   themes,   the   scale   of   surveys   and   map   restitution,   i.e.      1:50   000.   This   is   not   a   cadastral scale   but   one   that   is   detailed   enough   (the   area   of   ​​the   smallest   mapped   units   is   about   1   ha)   to   model   the   regional   ecosystem,   an essential   step   forward   in   the   process   leading   to   a   better   understanding   of   spatial   issues   for   the   use   and   appropriation   of   resources   in the Delta. 1 - Field surveys We   are   giving   here   only   the   minimum   number   of   indications   to   understand   the   methodology   we   implemented;   for   further details,   the   reader   will   refer   to   the   work   of   Hiernaux   et   al.    (cited   in   the   bibliography).   The   assessment   of   the   fodder   resources   of   the Delta   is   established   on   the   basis   of   a   phyto-ecological   study.   The   fodder   productivity,   floristic   composition,   sensitivity   to   grazing, etc.,   are   reported   in   relation   to   about   twenty   types   of   rangelands   defined   in   a   prior   phyto-ecological   study.   The   rangeland   types   are defined   both   by   the   characteristics   of   their   vegetation   –   floristic   composition   and   bio-morphological   structure   –   and   by   those   of   the environment– topographic and geomorphological positions, soil textures, flooding regime, mode of pastoral exploitation... The   phyto-ecological   analysis   was   carried   out   on   169   sites   of   100   m²   each,   127   located   in   the   flood   plains,   8   in   the   irrigated   rice fields   of   the   Office   du   Niger   and   34   others   in   the   unflooded   uplands   of   the   Delta.   The   implementation   of   a   methodical   and simultaneous   survey   of   the   characteristics   of   the   vegetation   and   the   environment   is   largely   inspired   by   the   method   developed   by CNRS researchers from the CEPE Louis Emberger Laboratory (M. Godron et al ., 1968, Ph.Daget et al .  1970). In   addition   to   the   plant   mass   measurements   accompanying   each   phyto-ecological   survey,   twelve   sites   were   devoted   to   vegetation production   measures   (a   mesh   enclosure   of   1000   to   1500   m²   was   set   for   each   site).   Measurements   and   treatments   were   carried   out within the enclosures, therefore on rangeland protected from grazing. The main measurements made within the enclosures were as follows: seasonal and inter-annual changes of standing mass and production when protected from grazing, effect of mowing and burning practiced at earlier or later stages, on grass regrowth, effect of several repetitive mowing regimes on the production of regrowth in the dry season, effect of several repetitive mowing regimes following initial burning on the production of regrowth in the dry season, trials in hay making, in terms of cutting period and mode of conservation. In   addition,   in   three   of   these   twelve   sites,   a   piece   of   rangeland   of   several   hectares   was   identified   and   monitored.   The   systematic monitoring   of   livestock   numbers   and   of   the   duration   of   their   presence   within   this   rangeland   was   used   to   estimate   seasonal   stocking rates.The   grass   mass   was   monitored   throughout   the   dry   season   in   the   grazed   rangeland   and   also   inside   mobile   cages   displaced   every 15 days to measure the regrowth of the grass under grazing. 2 - Data processing As   for   the   survey   method,   the   analysis   of   the   data   uses   the   calculation   procedure   recommended   by   the   researchers   of   the   Center of   Phytosociological   and   Ecological   Studies   Louis   Emberger   (CNRS   Montpellier)   where   the   calculations   were   carried   out.   Two approaches   were   combined:   one   more   analytical,   with   the   establishment   of   the   ecological   profiles   of   the   species   for   the   main ecological   variables,   the   other   more   synthetic   for   the   factorial   analyses   of   the   correspondences   which   are   carried   out   in   the   "species- site" and the "species-states of variables" matrices. For   each   plant   association,   flora,   ecology,   annual   and   seasonal   vegetation   and   fodder   production   are   characterized   on   the   basis   of systematic   information   collected   in   169   sites   (393   taxa   and   119   ecological   variables).   Statistical   analyses   of   these   data   established   the floristic   and   ecological   profiles   of   each   plant   combination.   These   are   so-called   'indexed   ecological   profiles’,   in   which   the   indication for   each   class   of   the   variable   or   taxon   considered   constitutes   a   threshold   of   probability   of   presence   or   absence   (of   the   plant association   in   the   situation   corresponding   to   a   class   of   the   variable,   or   of   the   species   in   the   plant   association).   The   method   known   as "indexed   ecological   profiles"   is   described   in   a   seminal   article   co-authored   by   B.   Gauthier,   M.   Godron,   P.   Hiernaux   and   J.   Lepart,   " A complementary type of ecological profile , Canadian Botanical Association, 1977, Vol 55, pp. 2859-2865. " It   consists   in   "analyzing   a   collection   of   phyto-ecological   surveys   taking   into   account   the   presence   of   plant   species   and   a   number   of variables   characterizing   the   environment".   Concretely,   a   test   assesses   the   sensitivity   of   the   plant   species,   or   of   the   plant   combination considered, to the state of the variable. Five cases of sensitivity are recorded and coded in the database. Table   1:   Codes   representing   the   intensity   of   species/plant   association   profile   linkages   or   plant   association/status   of variables.
· · · · · Cod e   Statistical significance   +++   The species is significantly  linked   to the state of the variable at  p <0.001   ++   The species is significantly  linked   to the state of the variable at  p <0.0 1  but not at p <0.001   +   The species is significantly  linked   to the state of the variable at  p <0.05but not at p <0.01      The species is  not  significantly  linked   to the state of the variable at  p <0.05   0   Sampling is not sufficient to conclude
The   identification   of   plant   combinations   and   field   checks   were   carried   out   between   September   1979   and   September   1982.   27 elementary   plant   combinations   were   recognized,   17   of   which   were   in   the   flood   plains   and   the   Delta   rice   fields,   3   in   the   northern lake plains and 7 on islands, banks and high plains.
*   MB   The   mosaic   of   the   banks   is   not,   strictly   speaking,   a   plant   combination,   but   a   complex   mosaic.   It provides   a   "shortcut"   for   different   flooded   formations   (VB,   B,   O,   VOR,   VSP,   VH,   AG,   ZB).   It   is   located   on   the edge   of   the   River’s   minor   bed,   on   the   spread   of   narrow   levees   and   channels   which   occupy   the   major   bed   of   the Niger,   the   Bani   and   their   main   tributaries.   Its   main   characteristics   are   the   subject   of   a   specific   sheet,   but   are   not determined by calculation like other mosaics.
Photo-interpretation and mapping The   mapping   was   done   on   the   basis   of   a   photo-interpretation   of   the   cover   75   MAL   32/500   23x23   cm   panchromatic   and   IRC   infra-red   color   plates   -   1:50   000   complemented   locally   by   covers   74   MAL   20/500   for   the   south-west   and   70/71   AO   891/500   for   the eastern   bank.   In   the   course   of   the   initial   surveys,   the   correspondence   between   the   nature   of   the   vegetation   type   and   that   of   the   topo- geomorphological   environment   was   established   on   a   case-by-case   basis,   on   the   basis   of   their   aspects   in   aerial   photographs.   Mark Haywood   then   proceeded   with   the   photo-interpretation   under   stereoscope   with   very   high   magnification,   delimitating   the   recognized vegetation   types.   The   boundaries   were   drawn   by   hand   on   the   topographic   map   of   the   National   Geographic   Institute   (I.G.N.)   and   the OICMA   map,   enlarged   by   the   kilometric   grid   method.   The   field   checks   were   carried   out   by   Pierre   Hiernaux,   Lassine   Diarra   and   Mark Haywood.   In      this   vegetation   map,   Mark   Haywood   refers   only   to   the   vegetation   types,   without   separating   the   cultivated   areas   whose location   and   extent   vary   from   year   to   year.   The   cultivated   areas   have   been   mapped   in   work   published   separately.   Indeed,   on   aerial photos,   particularly   infrared   ones,   one   can   "read"   the   vegetation   types   under   the   cultivated   plots   which,   even   if   they   are   used   for   rice- crops,   are   a   system   of   temporary   cultivation   alternating   with   fallows.   This   particularity   will   be   very   valuable   to   determine   which     vegetation   types   have   been   cleared   at   various   times.   However,   such   readings   are   not   possible   in   the   case   of   controlled   irrigated schemes   from   which   all   traces   of   previous   vegetation   types   have   disappeared.   Such   is   the   case   of   the   Office   du   Niger   schemes   which are   coded   R,   like   a   particular   plant   association.   BP,   B   and   VB   constitute   very   deep,   plain   “bourgoutières”   and   “Low   Vetiver   stands”, PAK   corresponds   to   very   deep   Vetiver   stands   with   Acacia   Kirkii    and   PAM   to   channels   and   low   plains   with   Mitragina   inernis .   OP,   O represent   deep   and   plain   wild   rice,   VOR   and   EOR,   deep   Vetiver   and   Eragrostis   savannas,   VSP   and   ESP,   medium   Vetiver   and Eragrostis   savannas,   VH   and   AC   the   high   plain   Vetiver   and   Eragrostis   savannas,   P   and   ZB   the   Panicaies   and   flood   edge   savannas. AG   is   a   shrubby   savanna   weakly   flooded   with   Andropogon   gayanus    and   the   plant   combinations   ranging   from   TA   to   TT   are   "dry" formations   that   normally   are   not   affected   by   flooding.   They   are   located   on   the   dry   margins   and   on   the   " togge ”   (sg.   toggere ),   the Fulani   name   for   the   never-flooded   mounds   in   the   Delta.   PAN,   PAS,   PAR   are   woody   savannas   extending   in   the   plains   where   the   flood is   very   irregular   and   related   to   local   run-offs   during   the   rainy   season   and   then,   at   the   end   of   the   year,   to   delayed   river   flooding     reverting   through   the   channels   toward   the   northern   lakes.   They   are   found   mainly   in   Farimaké.   A   map   unit   represents   one   of   the   27 elementary   plant   combinations   (28   with   the   mosaic   MB)   indicated   by   its   acronym   or,   more   often,   a   mosaic   of   two   elementary   types representing   a   gradient   along   a   slope   or   small   ripples   of   land   that   reflect   small   but   significant   variations   in   flood   conditions   over   a small   area.   In   some   cases,   photo-interpretation   would   have   made   it   possible   to   separate   the   constituent   types   from   the   mosaics,   but   the small   size   of   the   units   would   have   rendered   the   map   very   difficult   to   read.   In   the   rest   of   the   text,   the   term   "vegetation   type"   is   a generic   term.   The   term   "plant   combination"   refers   to   the   27   elementary   plant   combinations   identified   by   Pierre   Hiernaux.   They   are always   identified   by   an   acronym   consisting   of   one   to   three   letters,   as   shown   in   the   previous   table.   The   phrase   "vegetation   mosaic" designates   a      composite   of   vegetation   types.   They   are   always   identified   by   an   acronym   that   is   itself   composite,   the   two   constituent elements   of   which   are   separated   by   a   slash.   Thus   O/VOR   is   a   mosaic   whose   constituent   elements   are   the   vegetation   associations   O and VOR. The methods for calculating these mosaics will be described in the section on vegetation databases. Finally,   the   information   for   each   plant   association   is   grouped   into   three   sections:   floristics,   ecology   and   production.   These three sections have been kept in the database architecture and correspond to tables.
Table 2 : Codes and names of the twenty seven vegetation associations of the Inland Delta of the Niger  
Codes   Name of the vegetation association    AC   Very shallow  flood, savanna,  Eragrostis   barteri ,  Andropogon   canali culatus   AG    Sandy soil ,   savanna,  Andropogon   gayanus   B    Deep flood, plain,  Echinochloa   s tagnina   BP    Very deep flood, plain,   Echinochloa   s tagnina ,  Vossia   cuspidata   EOR    Mid flood, grassland,    Eragrostis   barteri   ESP    Shallow flood, grassland,  Eragrostis   barteri ,  Setaria anceps   MB *   Stream bead mosaic of vegetation associations   O    Mid flood, wild rice plain,  Oryza longistaminata,   Setaria anceps   OP    Deep flood wild rice,  Oryza longistaminata, Eleocharis dulcis   P    Marginal flood,  grassland, Panicum   anabaptisum   PAK    Very deep flood, wood l and,  Acacia  k irkii   PAM    Deep flood, thicket,  Mitragyna   inernis   PAN    Mid lacustrine flood, woodland,  Acacia  nilotica   PAR    Sh allow  lacust r ine flood, woodland,  Acacia raddiana     PAS    Slight   lacustrine flood, woodland,  Acacia seyal   R    Irrigated rice fields,  Oriza   sativa   TA    Upland, savanna,  Andropogon   g ayanus ,  Pilio s tigma reticulat um   TB    Palm grove,  Borassus  aethiopum   TC    Upland, anthropic woody savanna,   Celtis   integrifolia, Borassus   aethiopum   TD    Upland, thicket,  Diospyros mespiliformis, Acacia pennata   THY    Palm grove,  Hyphaene   thebaica   TS    Upland, shrub b y savanna,  Acacia seyal ,  Acacia sieberiana   TT    Upland, shrub b y   savanna,   Terminalia   macroptera   VB    Very deep flood, savanna,  Vetiveria   nigritiana ,  Vossia   cuspidata   VH    Very shallow   flood, savanna,  Vetiveria   nigritiana   VOR    Mid flood, savanna,  Vetiveria   nigritiana ,  Oriza   longistaminata   VSP    Shallow flood, savanna,  Vetiveria nigritiana ,  Hyparrhenia   rufa   ZB    Edges of highest floods,  Bergia suffruticosa
Field survey methods and data processing