Greenery seasonal dynamics is more crucial for monitoring urban environment. We have attempted to understand the quarterly variability of NDVI using Landsat data to get more accurate information about the vegetation effects on urban land surface temperature reduction. Along this, NDVI varied from year to year in cities. A similar trend was globally observed with a decreasing trend from 2000 to 2011 and an increasing trend from 2011 to 2020. This finding can be explained by the national commitments to vegetate cities with the governmental project “10 million people – 10 million trees” instituted in 2013 (MCVDD, 2019). Therefore, this governmental project would awake spirit in good management and safeguard of the existing urban vegetation in these cities. Likewise, the project goal was that each inhabitant must plant at least one tree per year with careful monitor. Moreover, the particular changes among cities could be the response to the vegetation density and the plant species diversity changes in these cities (Kinyanjui, 2011; Zohoun et al., 2020; Sehoun et al., 2021). The land surface temperature (LST) followed similar trend for the combined cities as well as for the cities taken individually. This result can be attributed to the changes in vegetation index. The land surface temperature was globally in decreasing (Fig. 3) with increasing NDVI (Fig. 2). The same findings were mentioned by Zhao et al. (2019) on factors affecting the decreased temperature. Likewise, a study in the horn of Africa has also found a negative relationship between air temperature and vegetation cover (Ghebrezgabher et al., 2020). However, although efforts were born to improve greenery in Beninese cities, a significant linear trend of LST was occurred in the city of Porto-Novo. We can justify this finding by the weakness in the management and monitoring of planted trees. It can be likely due to the density of built area (lack of areas for trees plantation), introduction of plant species without resident contentment (Osseni et al., 2014). Therefore, the NDVI trend was also somehow degraded in this city (Fig. 2).
A very significant difference was shown between seasonal variations of vegetation index (NDVI). This finding can be justified by season-wise vegetation statuses in the cities. Therefore, not only urban vegetation state varied according to space but also it varied with respect to climatic period (Malik et al., 2020). For combined cities as for individual studied cities, the period of January-March had the significant low mean NDVI value. This result can be similar to Piao et al. (2019), who found a low value for the period of June-July on the city of Nanjing (China). Then, the authors explained this difference by the interference of cloud and rain occurring. Other study found that during January to April in India, the vegetation showed very low NDVI because during this period very few leaves were available to reflect NDVI signature (Malik et al., 2020). The periods from January to March and April-June respectively in south and centre Benin, refer to dry season when overall vegetation features are in water stress (Ahokpossi, 2018; Oussou et al., 2022). Moreover, a link could be also made between the vegetation index and the vegetation phenology (Su et al., 2020). Likewise, it was also demonstrated that the seasonal variability is more important to describe the variation in phenological expression of urban vegetation (Li et al., 2017; Piao et al., 2019). Thereby, Su et al. (2020) went further to make relationship between season period and canopy phenology that are very determining in vegetation index importance. Along this, what should be the impacts of both seasonal and NDVI variability on land surface temperature in the study areas?
Annual mean LST builds a moderate negative correlation with NDVI in the combined cities as well as in separate cities. The similar result was found by many researchers who demonstrated an inverse correlation relationship between LST and NDVI whatever the land use types and polygons (Yue et al., 2007; Richards et al., 2020). But although the cooling effects that greening can procure to urban areas were widely demonstrated, the vegetation physiology due to environmental conditions can have strong impacts on this service (Norton et al., 2015). Thereby, the assessing seasonal variability in relationship between LST and NDVI, revealed a moderate positive correlation between LST and NDVI in the periods of January to March and April to June respectively for Porto-Novo (Figs. 5) and Parakou (Fig. 6). This finding expressed that the increase in urban vegetation in these cities respectively for these above periods could not decrease the LST. We can justify this finding by the occurrence of dry season at these periods respectively in Porto-Novo and in Parakou. Therefore, the vegetation that would be experiencing water stress cannot have great rates of evapotranspiration, then can lead higher surface temperature (Norton et al., 2015). Porto-Novo and Parakou usually experience respectively strong dry period with high water stress from January to March (Ahokpossi, 2018) and April to June (Ahokpossi, 2018; Lanmandjèkpogni et al., 2018). So, a wise investment in water availability such as urban greening irrigation would be more adequate for ensuring a sustainable urban surface temperature mitigation as well as other ecosystem service providing (Norton et al., 2015).