ASSESSING VEGETATION STRUCTURAL CHANGES IN OASIS AGRO-ECOSYSTEMS USING SENTINEL-2 IMAGE TIME SERIES : CASE STUDY FOR DRÂA-TAFILALET REGION MOROCCO

Nowadays, Moroccan oasis agro-ecosystems are under intense effect of natural and anthropogenic factors. Therefore, this essay proposes to use Leaf Area Index (LAI) to assess the consequences of the oases long-term biodegradation. The index was used as a widely-applied parameter of vegetation structure and an important indicator of plant growth and health. Therefore, a new optical multispectral Sentinel-2 data were used to build a long term LAI time series for the area of the Erfoud and Rissani oases, Errachidia province in Drâa-Tafilalet region in Morocco. Nine images of LAI spatial distribution over the study area were obtained by means of SNAP Biophysical Processor over the period since July 2015 till May 2018. Time series analysis of the resultedmaps has revealed a stable trend towards the average LAI decreasing and vegetation structure simplification as a consequence.


INTRODUCTION
Oasis agro-ecosystems are of exceptional value for Moroccan natural heritage.In particular, covering about 15 percent of the country, these biodiversity hotspots provide crucial economic, ecological, social and cultural goods and services.However, numerous factors threat to the oases survival.Among them are the abandonment of traditional cultivation and farming systems that have nurtured the ecosystems for centuries.Secondly, increased pressure on land and water resources and growing urbanization support the negative effect.Moreover, desertification is an additional natural effect due to the Sahara intense expansion.(Bodart, et al;,2005) Nowadays, the restoration of a sustainable oasis ecosystem is one of the main goals of the Moroccan Government and objective of international projects such as Revitalising Oasis Agro-ecosystems through a Sustainable, Integrated and Landscape Approach in the Draâ-Tafilalet Region (OASIL) (FAO and the GEF, 2017).Development of monitoring and evaluation indicators of oasis vegetation condition has a crucial significance for them.
Leaf area index (LAI) is considered an important crop parameter and an indicator of plant growth and health.It measures the amount of leaf material in an ecosystem and correlates directly with canopy foliage content and canopy structure (Borg, 2012).LAI is widely used for above-ground biomass assessment (Heiskanen, 2006), crop growth monitoring and yield estimation (Guindin-Garcia, 2012).
A broad variety of direct and indirect methods for remote estimation of the index has already proposed (Zheng, 2009 andMunier et al., 2018).Optical remote sensing methods build relationships exploiting in situ LAI measurements and/or as outputs of physical canopy radiative transfer models (Darvishzadeh, 2008).Novel European optical Sentinel-2 satellite provides perceptible advantages of spatial, spectral and temporal resolution for LAI estimation (Clevers, 2017).Moreover, European Space Agency trough Sentinels application platform (SNAP) proposes the processor that computing five main biophysical variables including LAI from Sentinel-2 reflectance (http://step.esa.int/main/toolboxes/sentinel-2toolbox/sentinel-2-toolbox-features/).Given all the above, the objective of the present study was threefold: i) to explore the capability of LAI product obtained from Sentinel-2 to evaluate vegetation structure in oasis agroecosystems at the different stages of seasonal development; ii) to assess interannual changes of vegetation structure in oasis agro-ecosystems; iii) to identify a main trend of vegetation structural changes in studied ecosystems for the defined period.

Study area
The study area included the Erfoud and Rissani oases of Errachidia province in Drâa-Tafilalet region.These oases comprise numerous species of tree crops, cereals, vegetables, fodder, in addition to aromatic, dye and medicinal species (FAO and the GEF, 2017).The structure of the oases are formed mainly by date palm cultivation and arable parcels.In many oases a three-layer structure has been arranged.There upper layer is formed by date palms which protect from excessive insolation the fruit trees (figs, almonds, olives, pomegranates etc.) and the arable crops (barley, wheat, sorghum, alfalfa and various vegetables) of the two lower layers.Interacting in this way the layers contribute to the formation of a mild microclimate.A well-managed palm grove is vital for the maintaining the traditional oasis structure (Garbati Pegna, 2017).
Sand encroachment threating especially inland and coastal provinces: endangered palm groves are estimated at 30,000 ha (80,000 inhabitants) in the province of Ouarzazate (Bodart, et al;, 2005)and 250,000 ha (200,000 inhabitants)in the province of Errachidiaspecies (FAO and the GEF, 2017).

Materials and Methods
For the aim of the study nine Sentinel-2 L1C images for a period from 2015-2018 were acquired from COPERNICUS Sentinels scientific data hub (https://scihub.copernicus.eu).The input Sentinel-2 images were radiometrically and atmospherically corrected using SNAP Sen2Cor processor (http://step.esa.int/main/third-party-plugins-2/sen2cor/).It performs the atmospheric-, terrain and cirrus correction of Top-Of-Atmosphere Level 1C input data.For each of preprocessed image LAIimage was derived by means of the Biophysical Processor.Its algorithm is based on specific radiative transfer models associated with strong assumptions, particularly regarding canopy architecture (turbid medium model) (Weiss, 2016).
Simultaneously, OSAVI (Optimized Soil-Adjusted Vegetation Index) was calculated and, applying the OSAVI threshold values, sparse vegetation of surrounding hills was masked out.This index was developed using the reflectance in the nearinfrared and red bands with an optimized soil adjustment coefficient (Rondeaux, 1996).OSAVI is robust to variability in soil brightness.It has enhanced sensitivity to vegetation cover greater than 50%.This index is best used in areas with relatively sparse vegetation where soil is visible through the canopy and where NDVI saturates (Ren, 2014;Aralova, 2015;Santin-Janin, et al., 2018).The obtained LAI distributions were under standard processing for time series analysis.The pixel-by-pixel simultaneous processing of all LAI maps resulted in spatial distributions of time series parameters, which describe the linear trend.These parameters are the all-time average and monthly rate.
The clearly expressed seasonal oscillations wereremoved from time series and the crisp linear trend was determined by the least squares method.

RESULTS
As a result, nine images of LAI spatial distribution over the study area were obtained.They are shown on Figure 4. Depending on vertical arrangement of leaf material that is corresponds to LAI values in the range from 0.2 to 5.0, seven classes of vegetation structure in the oasis agro-ecosystems were determined(Table 1 A stable trend towards the average LAI decreasing was detected as a result of the time series analysis of the index multitemporal spatial distributions (Figure 5).Interannual development of LAI that embrace all seven classes together with a linear trend to LAI decreasing illustrate changes of vegetated areas andare showed in Figure 6.From visual inspection of the results and comparison with highresolution imagery, the overall quality of class extraction seems to be satisfactory.Nevertheless, quantitative accuracy assessment is strongly required and should be performed in further research.

Figure 1 .b
Figure 1.Sentinel-2 MSI 10m resolution image of the area of interest.Data of image acquisition is July 16, 2015.Bands combination: 842 nm, 665 nm, 560 nm During each growing season, the dates of the images acquisition were defined according to annual distributions of average monthly temperature and precipitations.They are mid-term of May and November as start/end of the season with medium temperature and significant precipitations, and July as the hottest and driest period (https://weather-andclimate.com/average-monthly-Rainfall-Temperature-Sunshine,erfoud-meknes-tafilalet-ma,Morocco ).

Figure 5 .
Figure 5.Total trend of LAI interannual changes within the study area over the period since July 2015 till May 2018.

Figure 6 .
Figure 6.Interannual development of LAI and a resulting linear trend of changes in total vegetation amount.

Table 1 .
). Spatial distributions of the within the study area over the period July 2015May 2018.Obtained LAI maps shows that medium value classes have replaced by low value classeseach season since 2015, while high value classes are critically small or disappeared.Thus, the maps describe the vegetation conditions dynamics and indicate vegetation structure simplification within the investigated oases.Areas of the distinguished seven classes over the period July 2015 toMay 2018 are presented in Table 1.Areas of LAI classes, ha The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-4/W12, 2019 5th International Conference on Geoinformation Science -GeoAdvances 2018, 10-11 October 2018, Casablanca, Morocco