@article{fdi:010061859,
  title = {{S}patio- temporal variability of vegetation cover over {M}orocco (1982-2008) : linkages with large scale climate and predictability},
  author = {{J}arlan, {L}ionel and {D}riouech, {F}. and {T}ourre, {Y}. and {D}uchemin, {B}eno{\^i}t and {B}ouyssie, {M}. and {A}baoui, {J}. and {O}uldbba, {A}. and {M}okssit, {A}. and {C}hehbouni, {A}bdelghani},
  editor = {},
  language = {{ENG}},
  abstract = {{T}he dominant patterns of vegetation cover interannual variability over {M}orocco are isolated using rotated extended empirical orthogonal functions applied to {AVHRR} {NDVI} data (1982-2008). {T}he three leading modes capture the {NDVI} signal at the vegetation peak for three distinct locations: mode 1 (18.7% of total variance) is located along the {A}tlantic coastline, mode 2 (13.1%) is southwest of the {R}iff {M}ountain whilst mode 3 (11.2%) is along the {M}editerranean coastline. {C}orrelations between the {NDVI} time coefficients for the modes {A}tlantic' and {M}editerranean' dominated by annuals and precipitation amount during the early stage of the vegetation cycle ({NDJ}) are found. {S}ignificant fluctuations of {NDVI} time coefficients are isolated: a quasi-biennial signal is present in the three modes and an additional quasi-quadriennial (approximate to 4.4 years) signal is identified for the {A}tlantic' mode only. {C}onnection between vegetation activity and atmospheric and oceanic climate signals are sought using time-lag correlation analyses. {T}he {NAO} during fall-beginning of winter ({NDJ}) is found to impact vegetation peak for the {A}tlantic' mode while the {S}candinavian {P}attern is related to {NDVI} peak over the {A}tlantic' and {R}iff' latter in the season ({DJF}). {A} significant connection is also found between vegetation over the {A}tlantic' mode and the {R}iff' and the {A}tlantic {N}ino' mode leading the {SST} variability in the equatorial {A}tlantic with a 6-months lag. {F}inally, linkages between {NDVI} and climate information are used to build a seasonal prediction model for {NDVI} using multiple linear regression. {T}he {NDVI} anomalies during {M}arch-{A}pril may be predicted with a reasonable accuracy from {J}anuary with 79% of explained variance, 60% and 72% for the {A}tlantic', the {R}iff' and the {M}editerranean' regions, respectively. {R}esults have (1) direct impacts for a better understanding of the role of large-scale climate signals on vegetation cover over {M}orocco and (2) contribute to the implementation of an agricultural early warning system.},
  keywords = {remote sensing ; vegetation ; {M}orocco ; {NDVI} ; {NAO} ; predictability ; {MAROC}},
  booktitle = {},
  journal = {{I}nternational {J}ournal of {C}limatology},
  volume = {34},
  numero = {4},
  pages = {1245--1261},
  ISSN = {0899-8418},
  year = {2014},
  DOI = {10.1002/joc.3762},
  URL = {https://www.documentation.ird.fr/hor/fdi:010061859},
}