An Exploration of Terrain Effects on Land Surface Phenology across the Qinghai–Tibet Plateau Using Landsat ETM+ and OLI Data

Drivers of contemporary and future changes in Arctic seasonal transition dates for a tundra site in coastal Greenland

Climate change has had a significant impact on the seasonal transition dates of Arctic tundra ecosystems, causing diverse variations between distinct land surface classes. However, the combined effect of multiple controls as well as their individual effects on these dates remains unclear at various scales and across diverse land surface classes. Here we quantified spatiotemporal variations of three seasonal transition dates (start of spring, maximum normalized difference vegetation index (NDVImax ) day, end of fall) for five dominating land surface classes in the ice-free Greenland. Using a distributed snow model, structural equation modeling, and a random forest model, based on ground observations and remote sensing data, we assessed the indirect and direct effects of climate, snow, and terrain on seasonal transition dates. We then presented new projections of likely changes in seasonal transition dates under six future climate scenarios. The coupled climate, snow cover, and terrain conditions explained up to 61% of seasonal transition dates across different land surface classes. Snow ending day played a crucial role in the start of spring and timing of NDVImax . A warmer June and a decline in wind could advance the NDVImax day. Increased precipitation and temperature during July-August are the most important for delaying the end of fall. We projected that a 1-4.5°C increase in temperature and a 5%-20% increase in precipitation would lengthen the spring-to-fall period for all five land surface classes by 2050, thus the current order of spring-to-fall lengths for the five land surface classes could undergo notable changes. Tall shrubs and fens would have a longer spring-to-fall period under the warmest and wettest scenario, suggesting a competitive advantage for these vegetation communities. This study's results illustrate controls on seasonal transition dates and portend potential changes in vegetation composition in the Arctic under climate change.

Effect of Altitude and Topography on Vegetation Phenological Changes in the Niubeiliang Nature Reserve of Qinling Mountains, China

Due to the fragility of the habitats in mountain nature reserves, the vegetation is extremely sensitive to climate change, and its phenological changes are more specific. Therefore, it is of great significance to study the effects of topography and climate on the vegetation phenology in mountain nature reserves. Based on the vegetation phenology data retrieved from MODIS EVI2 during 2000 to 2017, combined with temperature data, spatial trend analysis and correlation analysis methods were used to study the effects of topographic and climatic factors on vegetation phenology in the Niubeiliang Nature Reserve of the Qinling Mountains. The results showed that the GSS (growing season start) was advanced with a rate of 4.24 days/10a, and the rates in the northern and southern slopes were almost the same; the GSE (growing season end) was delayed with a rate of 3.29 days/10a, and the rate in the northern slope was faster; and the GSL (growing season length) was prolonged. There were seasonal differences and north–south differences in the effects of topography on the phenophase. The phenophase changed regularly with the increase in altitude. The higher the altitude, the more significant the effect. The steeper the slope, the later the GSS, the earlier the GSE, and the more significant its effect on the GSE. The aspect had little effect on GSS but a more significant effect on GSE, which was the latest on the sunny slope and the earliest on the zero slope. Temperature affected both the GSS advance and the GSE delay, and both had a time-lag effect of approximately 2–3 months. Its effect was more significant in the GSE, in the southern slopes, and in the high-altitude areas.

Reconstructing High-Spatiotemporal-Resolution (30 m and 8-Days) NDVI Time-Series Data for the Qinghai–Tibetan Plateau from 2000–2020

As the largest and highest alpine ecoregion in the world, the Qinghai–Tibetan Plateau (QTP) is extremely sensitive to climate change and has experienced extraordinary warming during the past several decades; this has greatly affected various ecosystem processes in this region such as vegetation production and phenological change. Therefore, numerous studies have investigated changes in vegetation dynamics on the QTP using the satellite-derived normalized-difference vegetation index (NDVI) time-series data provided by the Moderate-Resolution Imaging Spectroradiometer (MODIS). However, the highest spatial resolution of only 250 m for the MODIS NDVI product cannot meet the requirement of vegetation monitoring in heterogeneous topographic areas. In this study, therefore, we generated an 8-day and 30 m resolution NDVI dataset from 2000 to 2020 for the QTP through the fusion of 30 m Landsat and 250 m MODIS NDVI time-series data. This dataset, referred to as QTP-NDVI30, was reconstructed by employing all available Landsat 5/7/8 images (>100,000 scenes) and using our recently developed gap-filling and Savitzky–Golay filtering (GF-SG) method. We improved the original GF-SG approach by incorporating a module to process snow contamination when applied to the QTP. QTP-NDVI30 was carefully evaluated in both quantitative assessments and visual inspections. Compared with reference Landsat images during the growing season in 100 randomly selected subregions across the QTP, the reconstructed 30 m NDVI images have an average mean absolute error (MAE) of 0.022 and a spatial structure similarity (SSIM) above 0.094. We compared QTP-NDVI30 with upscaled cloud-free PlanetScope images in some topographic areas and observed consistent spatial variations in NDVI between them (averaged SSIM = 0.874). We further examined an application of QTP-NDVI30 to detect vegetation green-up dates (GUDs) and found that QTP-NDVI30-derived GUD data show general agreement in spatial patterns with the 250 m MODIS GUD data, but provide richer spatial details (e.g., GUD variations at the subpixel scale). QTP-NDVI30 provides an opportunity to monitor vegetation and investigate land-surface processes in the QTP region at fine spatiotemporal scales.

Satellite-Based Monitoring on Green-Up Date for Optimizing the Rest-Grazing Period in Xilin Gol Grassland

Ecological degradation has occurred in global grasslands and has impaired their ecosystem services severely, so ecological conservation of grasslands should be focused more on the effectiveness of management measures. The green-up process decides the year-round forage yield and ecological conditions of grasslands. Adopting rest-grazing during the green-up process can guarantee a successful green-up, thus realizing more economic benefits without grassland degradation. Therefore, studies should pay more attention to whether the green-up process is really covered by the rest-grazing period or not. We analyze the spatiotemporal variations and the stability of the annual green-up date in Xilin Gol Grassland from 2000 to 2018 based on MODIS time series images and compare the green-up date with the rest-grazing period to assess the effectiveness of the rest-grazing policy. The results show that the green-up date of Xilin Gol Grassland had advanced 15 days on average because of the increasing trend of both temperature and precipitation during 2000~2018. The green-up date is mostly 120~130 d in the east, about 10 days earlier than the west (130~140 d) and 20 days earlier than in the central areas (140~150 d), also because of the spatial variations of temperature and precipitation. The coefficient of variation (CV) of the green-up date showed a significant negative correlation with precipitation, so the green-up date is more unstable in the arid areas. The rest-grazing period started more than 45 days earlier than the green-up date and failed to cover it in several years, which occurred more frequently in southern counties. The average green-up date appeared after rest-grazing started in over 98% of areas, and the time gap is 15~45 days in 88% of areas, which not only could not avoid grassland degradation effectively but also increased herdsmen’s life burden. This study aims to accurately grasp the temporal and spatial variations of the green-up date in order to provide references for adjusting a more proper rest-grazing period, thus promoting ecological conservation and sustainable development of animal husbandry.

Characteristics of Greening along Altitudinal Gradients on the Qinghai–Tibet Plateau Based on Time-Series Landsat Images

The Qinghai–Tibet Plateau (QTP) is ecologically fragile and is especially sensitive to climate change. Previous studies have shown that the vegetation on the QTP is undergoing overall greening with variations along altitudinal gradients. However, the mechanisms that cause the differences in the spatiotemporal patterns of vegetation greening among different types of terrain and vegetation have not received sufficient attention. Therefore, in this study, we used a Landsat NDVI time-series for the period 1992–2020 and climate data to observe the effects of terrain and vegetation types on the spatiotemporal patterns in vegetation greening on the QTP and to analyze the factors driving this greening using the geographical detector and the velocity of the vertical movement of vegetation greenness isolines. The results showed the following: (1) The vertical movement of the vegetation greenness isolines was affected by the temperature and precipitation at all elevations. The precipitation had a more substantial effect than the temperature below 3000 m. In contrast, above 3000 m, the temperature had a greater effect than the precipitation. (2) The velocity of the vertical movement of the vegetation greenness isolines of woody plants was higher than that of herbaceous plants. (3) The influence of slope on the vertical movement of vegetation greenness isolines was more significant than that of the aspect. The results of this study provided details of the spatiotemporal differences in vegetation greening between different types of terrain and vegetation at a 30-m scale as well as of the underlying factors driving this greening. These results will help to support ecological protection policies on the QTP.

Land Surface Phenology Retrieval through Spectral and Angular Harmonization of Landsat-8, Sentinel-2 and Gaofen-1 Data

Land Surface Phenology is an important characteristic of vegetation, which can be informative of its response to climate change. However, satellite-based identification of vegetation transition dates is hindered by inconsistencies in different observation platforms, including band settings, viewing angles, and scale effects. Therefore, time-series data with high consistency are necessary for monitoring vegetation phenology. This study proposes a data harmonization approach that involves band conversion and bidirectional reflectance distribution function (BRDF) correction to create normalized reflectance from Landsat-8, Sentinel-2A, and Gaofen-1 (GF-1) satellite data, characterized by the same spectral and illumination-viewing angles as the Moderate-Resolution Imaging Spectroradiometer (MODIS) and Nadir BRDF Adjusted Reflectance (NBAR). The harmonized data are then subjected to the spatial and temporal adaptive reflectance fusion model (STARFM) to produce time-series data with high spatio–temporal resolution. Finally, the transition date of typical vegetation was estimated using regular 30 m spatial resolution data. The results show that the data harmonization method proposed in this study assists in improving the consistency of different observations under different viewing angles. The fusion result of STARFM was improved after eliminating differences in the input data, and the accuracy of the remote-sensing-based vegetation transition date was improved by the fused time-series curve with the input of harmonized data. The root mean square error (RMSE) estimation of the vegetation transition date decreased by 9.58 days. We concluded that data harmonization eliminates the viewing-angle effect and is essential for time-series vegetation monitoring through improved data fusion.

How Does Spring Phenology Respond to Climate Change in Ecologically Fragile Grassland? A Case Study from the Northeast Qinghai-Tibet Plateau

Vegetation phenology is an important indicator of global climate change, and the response of grassland phenology to climate change is particularly sensitive in ecologically fragile areas. To enhance the ecological security of the Tibetan Plateau, it is crucial to determine the relationship between fluctuations in the start of the growing season (SOS) and the response to environmental factors. We investigated the trends of the intra-annual (ten-day) and interannual spatiotemporal dynamics of the SOS on the Northeast Qinghai-Tibet Plateau (NQTP) from 2000–2020 with MOD09GA data. We identified the response relationships with environmental factors (climate, terrain) using the maximum value composite method and the Savitzky–Golay filtering and dynamic threshold method. The SOS was concentrated from the 110th to 150th days; the average annual SOS was on the 128th day, with a spatial pattern of “early in the east and late in the west”. The overall trend of the SOS was advanced (45.48%); the regions with the advanced trend were mainly distributed in the eastern part of the NQTP. The regions with a delayed SOS were mainly concentrated in the higher-altitude regions in the southwest (38.31%). The temperature, precipitation and SOS exhibited a reverse fluctuation trend around the midpoint of 2010. Precipitation affected the SOS earlier than temperature. When temperature became a limitation of the SOS, precipitation had a more significant regulatory effect on the SOS. The SOS and aspect, slope and altitude were distributed in axisymmetric, pyramidal and inverted pyramidal shapes, respectively. The SOS on shaded slopes was earlier and more intensive than that on sunny slopes. With increasing slope, the area of the SOS decreased, and it occurred later. The SOS area was largest at 4500–5000 m and decreased at lower and higher altitude intervals. The SOS occurred later as altitude increased.

Populus euphratica Phenology and Its Response to Climate Change in the Upper Tarim River Basin, NW China

Quantifying the phenological variations of Populus euphratica Olivier (P. euphratica) resulting from climate change is vital for desert ecosystems. There has previously been great progress in the influence of climate change on vegetation phenology, but knowledge of the variations in P. euphratica phenology is lacking in extremely arid areas. In this study, a modified method was proposed to explore P. euphratica phenology and its response to climate change using 18-year Global Land Surface Satellite (GLASS) leaf area index (LAI) time series data (2000–2017) in the upper Tarim River basin. The start of the growing season (SOS), length of the growing season (LOS), and end of the growing season (EOS) were obtained with the dynamic threshold method from the reconstructed growth time series curve by using the Savitzky–Golay filtering method. The grey relational analysis (GRA) method was utilized to analyze the influence between the phenology and the key climatic periods and factors. Importantly, we also revealed the positive and negative effects between interannual climate factors and P. euphratica phenology using the canonical correlation analysis (CCA) method, and the interaction between the SOS in spring and EOS in autumn. The results revealed that trends of P. euphratica phenology (i.e., SOS, EOS, and LOS) were not significant during the period from 2000–2017. The spring temperature and sunshine duration (SD) controlled the SOS, and the EOS was mainly affected by the temperature and SD from June–November, although the impacts of average relative humidity (RH) and precipitation (PR) on the SOS and EOS cannot be overlooked. Global warming may lead to SOS advance and EOS delay, and the increase in SD and PR may lead to earlier SOS and later EOS. Runoff was found to be a more key factor for controlling P. euphratica phenology than PR in this region.

Deep Learning-Based Phenological Event Modeling for Classification of Crops

Classification of crops using time-series vegetation index (VI) curves requires appropriate modeling of phenological events and their characteristics. The current study explores the use of capsules, a group of neurons having an activation vector, to learn the characteristic features of the phenological curves. In addition, joint optimization of denoising and classification is adopted to improve the generalizability of the approach and to make it resilient to noise. The proposed approach employs reconstruction loss as a regularizer for classification, whereas the crop-type label is used as prior information for denoising. The activity vector of the class capsule is applied to sample the latent space conditioned on the cell state of a Long Short-Term Memory (LSTM) that integrates the sequences of the phenological events. Learning of significant phenological characteristics is facilitated by adversarial variational encoding in conjunction with constraints to regulate latent representations and embed label information. The proposed architecture, called the variational capsule network (VCapsNet), significantly improves the classification and denoising results. The performance of VCapsNet can be attributed to the suitable modeling of phenological events and the resilience to outliers and noise. The maxpooling-based capsule implementation yields better results, particularly with limited training samples, compared to the conventional implementations. In addition to the confusion matrix-based accuracy measures, this study illustrates the use of interpretability-based evaluation measures. Moreover, the proposed approach is less sensitive to noise and yields good results, even at shallower depths, compared to the main existing approaches. The performance of VCapsNet in accurately classifying wheat and barley crops indicates that the approach addresses the issues in crop-type classification. The approach is generic and effectively models the crop-specific phenological features and events. The interpretability-based evaluation measures further indicate that the approach successfully identifies the crop transitions, in addition to the planting, heading, and harvesting dates. Due to its effectiveness in crop-type classification, the proposed approach is applicable to acreage estimation and other applications in different scales.

Deriving Annual Double-Season Cropland Phenology Using Landsat Imagery

Cropland phenology provides key information in managing agricultural practices and modelling crop yield. However, most of the existing phenological products have coarse spatial resolution ranging from 250 to 8000 m, which is not sufficient to capture the critical spatial details of cropland phenology at the landscape scale. Landsat imagery provides an unprecedented data source to generate 30-m spatial resolution phenological products. This paper explored the potential of utilizing multi-year Landsat enhanced vegetation index to derive annual phenological metrics of a double-season agricultural land from 1993 to 2009 in a sub-urban area of Shanghai, China. We used all available Landsat TM and ETM+ observations (538 scenes) and developed a Landsat double-cropping phenology (LDCP) algorithm. LDCP captures the temporal trajectory of multi-year enhanced vegetation index time series very well, with the degree of fitness ranging from 0.78 to 0.88 over the study regions. We found good agreements between derived annual phenological metrics and in situ observation, with root mean square error ranging from 8.74 to 18.04 days, indicating that the proposed LDCP is capable of detecting double-season cropland phenology. LDCP could reveal the spatial heterogeneity of cropland phenology at parcel scales. Phenology metrics were retrieved for approximately one-third and two-thirds of the 17 years for the first and second cropping cycles, respectively, depending on the number of good quality Landsat data. In addition, we found an advanced peak of season for both cropping cycles in 50–60% of the study area, and a delayed start of season for the second cropping cycle in 50–70% of the same area. The potential drivers of those trends might be climate warming and changes in agricultural practices. The derived cropland phenology can be used to help estimate historical crop yields at Landsat spatial resolution, providing insights on evaluating the effects of climate change on temporal variations of crop growth, and contributing to food security policy making.