Assessment of Vegetation Phenological Extractions Derived From Three Satellite-Derived Vegetation Indices Based on Different Extraction Algorithms Over the Tibetan Plateau

Remote sensing phenology retrieval can remedy the deficiencies in field investigations and has the advantage of catching the continuous characteristics of phenology on a large scale. However, there are some discrepancies in the results of remote sensing phenological metrics derived from different vegetation indices based on different extraction algorithms, and there are few studies that evaluate the impact of different vegetation indices on phenological metrics extraction. In this study, three satellite-derived vegetation indices (enhanced vegetation index, EVI; normalized difference vegetation index, NDVI; and normalized difference phenology index, NDPI; calculated using surface reflectance data from MOD09A1) and two algorithms were used to detect the start and end of growing season (SOS and EOS, respectively) in the Tibetan Plateau (TP). Then, the retrieved SOS and EOS were evaluated from different aspects. Results showed that the missing rates of both SOS and EOS based on the Seasonal Trend Decomposition by LOESS (STL) trendline crossing method were higher than those based on the seasonal amplitude method (SA), and the missing rate varied using different vegetation indices among different vegetation types. Also, the temporal and spatial stabilities of phenological metrics based on SA using EVI or NDPI were more stable than those from others. The accuracy assessment based on ground observations showed that phenological metrics based on SA had better agreements with ground observations than those based on STL, and EVI or NDVI may be more appropriate for monitoring SOS than NDPI in the TP, while EOS from NDPI had better agreements with ground-observed EOS. Besides, the phenological metrics over the complex terrain also presented worse performances than those over the flat terrain. Our findings suggest that previous results of inter-annual variability of phenology from a single data or method should be treated with caution.

Forest cover trend analysis using MODIS time series and its climatic responses in the Mari El Republic of Russia

This study assesses whether MODIS NDVI satellite data time series can be used to detect changes in forest phenology over the different forest types of the Mari El Republic of Russia. Due to the severe climatic conditions, coniferous and deciduous forests of this region are especially vulnerable to climate change, which can lead to stresses from droughts and increase the frequency of wild fires in the long term. Time series analysis was applied to 16-day composite MODIS (MOD13Q1) (250 m) satellite data records (2000-2020) for the investigated territory, based on understanding that the NDVI trend vectors would enable detection of phenological changes in forest cover. There was also the determination of land cover/land use change for the area and examination of meteorological data for the investigated period. For the study, we utilized four phenological metrics: start of season (SOS), end of season (EOS), length of season (LOS), and Maximum vegetation index (MVI). The NDVI MODIS data series were smoothed in the TimeSAT software using the Savitsky-Golay filter. The results of the study show that over the 20-years period variations in phenological metrics do not have a significant impact on the productivity and growth of forest ecosystems in the Mari El Republic.

Validation of Sentinel-2, MODIS, CGLS, SAF, GLASS and C3S Leaf Area Index Products in Maize Crops

We proposed a direct approach to validate hectometric and kilometric resolution leaf area index (LAI) products that involved the scaling up of field-measured LAI via the validation and recalibration of the decametric Sentinel-2 LAI product. We applied it over a test study area of maize crops in northern China using continuous field measurements of LAINet along the year 2019. Sentinel-2 LAI showed an overall accuracy of 0.67 in terms of Root Mean Square Error (RMSE) and it was used, after recalibration, as a benchmark to validate six coarse resolution LAI products: MODIS, Copernicus Global Land Service 1 km Version 2 (called GEOV2) and 300 m (GEOV3), Satellite Application Facility EUMETSAT Polar System (SAF EPS) 1.1 km, Global LAnd Surface Satellite (GLASS) 500 m and Copernicus Climate Change Service (C3S) 1 km V2. GEOV2, GEOV3 and MODIS showed a good agreement with reference LAI in terms of magnitude (RMSE ≤ 0.29) and phenology. SAF EPS (RMSE = 0.68) and C3S V2 (RMSE = 0.41), on the opposite, systematically underestimated high LAI values and showed systematic differences for phenological metrics: a delay of 6 days (d), 20 d and 24 d for the start, peak and the end of growing season, respectively, for SAF EPS and an advance of −4 d, −6 d and −6 d for C3S.

National-Scale Cropland Mapping Based on Phenological Metrics, Environmental Covariates, and Machine Learning on Google Earth Engine

Many challenges prevail in cropland mapping over large areas, including dealing with massive volumes of datasets and computing capabilities. Accordingly, new opportunities have been opened at a breakneck pace with the launch of new satellites, the continuous improvements in data retrieval technology, and the upsurge of cloud computing solutions such as Google Earth Engine (GEE). Therefore, the present work is an attempt to automate the extraction of multi-year (2016–2020) cropland phenological metrics on GEE and use them as inputs with environmental covariates in a trained machine-learning model to generate high-resolution cropland and crop field-probabilities maps in Morocco. The comparison of our phenological retrievals against the MODIS phenology product shows very close agreement, implying that the suggested approach accurately captures crop phenology dynamics, which allows better cropland classification. The entire country is mapped using a large volume of reference samples collected and labelled with a visual interpretation of high-resolution imagery on Collect-Earth-Online, an online platform for systematically collecting geospatial data. The cropland classification product for the nominal year 2019–2020 showed an overall accuracy of 97.86% with a Kappa of 0.95. When compared to Morocco’s utilized agricultural land (SAU) areas, the cropland probabilities maps demonstrated the ability to accurately estimate sub-national SAU areas with an R-value of 0.9. Furthermore, analyzing cropland dynamics reveals a dramatic decrease in the 2019–2020 season by 2% since the 2018–2019 season and by 5% between 2016 and 2020, which is partly driven by climate conditions, but even more so by the novel coronavirus disease 2019 (COVID-19) that impacted the planting and managing of crops due to government measures taken at the national level, like complete lockdown. Such a result proves how much these methods and associated maps are critical for scientific studies and decision-making related to food security and agriculture.

Characterizing bracken fern phenological cycle using time series data derived from Sentinel-2 satellite sensor

Bracken fern is an invasive plant that has caused serious disturbances in many ecosystems due to its ability to encroach into new areas swiftly. Adequate knowledge of the phenological cycle of bracken fern is required to serve as an important tool in formulating management plans to control the spread of the fern. This study aimed to characterize the phenological cycle of bracken fern using NDVI and EVI2 time series data derived from Sentinel-2 sensor. The TIMESAT program was used for removing low quality data values, model fitting and for extracting bracken fern phenological metrics. The Sentinel-2 satellite-derived phenological metrics were compared with the corresponding bracken fern phenological events observed on the ground. Findings from our study revealed that bracken fern phenological metrics estimated from satellite data were in close agreement with ground observed phenological events with R2 values ranging from 0.53–0.85 (p < 0.05). Although they are comparable, our study shows that NDVI and EVI2 differ in their ability to track the phenological cycle of bracken fern. Overall, EVI2 performed better in estimating bracken fern phenological metrics as it related more to ground observed phenological events compared to NDVI. The key phenological metrics extracted in this study are critical for improving the precision in the controlling of the spread of bracken fern as well as in implementing active protection strategies against the invasion of highly susceptible rangelands.

Evaluation of Vegetation Indices and Phenological Metrics Using Time-Series MODIS Data for Monitoring Vegetation Change in Punjab, Pakistan

In arid and semi-arid regions, it is essential to monitor the spatiotemporal variability and dynamics of vegetation. Among other provinces of Pakistan, Punjab has produced a significant number of crops. Recently, Punjab, Pakistan, has been described as a global hotspot for extremes of climate change. In this study, the soil adjusted vegetation index (SAVI), normalized vegetation difference index (NDVI), and enhanced vegetation index (EVI) were comprehensively evaluated to monitor vegetation change in Punjab, Pakistan. The time-series MODIS (Moderate Resolution Imaging Spectroradiometer) data of different periods were used. The mean annual variability of the above vegetation indices (VIs) from 2000 to 2019 was evaluated and analyzed. For each type of vegetation, two phenological metrics (i.e., for the start of the season and end of the season) were calculated and compared. The spatio-temporal image analysis of the mean annual vegetation indices revealed similar patterns and varying vegetation conditions. In the forests and vegetation areas with sparse vegetation, the EVI showed high uncertainty. The phenological metrics of all vegetation indices were consistent for most types of vegetation. However, the NDVI result had the greatest variance between the start and end of season. The lowest annual VI variability was mainly observed in the southern part of the study area (less than 10% of the study area) based on the statistical analysis of spatial variability. The mean annual spatial variability of NDVI was <20%, SAVI was 30%, and EVI ranged between 10–20%. More than 40% of the variability was observed in the NDVI and SAVI vegetation indices.

Applicability of Smoothing Techniques in Generation of Phenological Metrics of Tectona grandis L. Using NDVI Time Series Data

Information on phenological metrics of individual plant species is meager. Phenological metrics generation for a specific plant species can prove beneficial if the species is ecologically or economically important. Teak, a dominating tree in most regions of the world has been focused on in the present study due to its multiple benefits. Forecasts on such species can attain a substantial improvement in their productivity. MODIS NDVI time series when subjected to statistical smoothing techniques exhibited good output with Tukey’s smoothing (TS) with a low RMSE of 0.042 compared to single exponential (SE) and double exponential (DE). Phenological metrics, namely, the start of the season (SOS), end of the season (EOS), maximum of the season (MAX), and length of the season (LOS) were generated using Tukey-smoothed MODIS NDVI data for the years 2003–2004 and 2013–2014. Post shifts in SOS and EOS by 14 and 37 days respectively with a preshift of 28 days in MAX were observed in the year 2013–2014. Preshift in MAX was accompanied by an increase in greenness exhibiting increased NDVI value.LOS increased by 24 days in the year 2013–2014, showing an increase in the duration of the season of teak. Dates of these satellite-retrieved phenological occurrences were validated with ground phenological data calculated using crown cover assessment. The present study demonstrated the potential of a spatial approach in the generation of phenometrics for an individual plant species, which is significant in determining productivity or a crucial trophic link for a given region.

Using Multi-Temporal Satellite Data to Analyse Phenological Responses of Rubber (Hevea brasiliensis) to Climatic Variations in South Sumatra, Indonesia

Land surface phenology derived from satellite data provides insights into vegetation responses to climate change. This method has overcome laborious and time-consuming manual ground observation methods. In this study, we assessed the influence of climate on phenological metrics of rubber (Hevea brasiliensis) in South Sumatra, Indonesia, between 2010 and 2019. We modelled rubber growth through the normalised difference vegetation index (NDVI), using eight-day surface reflectance images at 250 m spatial resolution, sourced from NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) Terra and Aqua satellites. The asymmetric Gaussian (AG) smoothing function was applied on the model in TIMESAT to extract three phenological metrics for each growing season: start of season (SOS), end of season (EOS), and length of season (LOS). We then analysed the effect of rainfall and temperature, which revealed that fluctuations in SOS and EOS are highly related to disturbances such as extreme rainfall and elevated temperature. Additionally, we observed inter-annual variations of SOS and EOS associated with rubber tree age and clonal variability within plantations. The 10-year monthly climate data showed a significant downward and upward trend for rainfall and temperature data, respectively. Temperature was identified as a significant factor modulating rubber phenology, where an increase in temperature of 1 °C advanced SOS by ~25 days and EOS by ~14 days. These results demonstrate the capability of remote sensing observations to monitor the effects of climate change on rubber phenology. This information can be used to improve rubber management by helping to identify critical timing for implementation of agronomic interventions.

Shifts in leaf senescence across the Northern Hemisphere in response to seasonal warming

Shifts in plant phenology under ongoing warming affect global vegetation dynamics and carbon assimilation of the biomes. The response of leaf senescence to climate is crucial for predicting changes in the physiological processes of trees at ecosystem scale. We used long-term ground observations, phenological metrics derived from PhenoCam, and satellite imagery of the Northern Hemisphere to show that the timings of leaf senescence can advance or delay in case of warming occurring at the beginning (before June) or during (after June) the main growing season, respectively. Flux data demonstrated that net photosynthetic carbon assimilation converted from positive to negative at the end of June. These findings suggest that leaf senescence is driven by carbon assimilation and nutrient resorption at different growth stages of leaves. Our results provide new insights into understanding and modelling autumn phenology and carbon cycling under warming scenarios.