Determination of phenological dates in boreal regions using normalized difference water index

Monitoring and understanding plant phenology are important in the context of studies of terrestrial productivity and global change. Vegetation phenology, such as dates of onsets of greening up and leaf senescence, has been determined by remote sensing using mainly the normalized difference vegetation index (NDVI). In boreal regions, the results suffer from significant uncertainties because of the effect of snow on NDVI. In this paper, SPOT VEGETATION S10 data over Siberia have been analysed to define a more appropriate method. The analysis of time series of NDVI, normalized difference snow index (NDSI), and normalized difference water index (NDWI), together with an analysis of in situ phenological records in Siberia, shows that the vegetation phenology can be detected using NDWI, with small effect of snow. In spring, the date of onset of greening up is taken as the date at which NDWI starts increasing, since NDWI decreases with snowmelt and increases with greening up. In the fall, the date of onset of leaf coloring is taken as the date at which NDWI starts decreasing, since NDWI decreases with senescence and increases with snow accumulation. The results are compared to the results obtained using NDVI-based methods, taking in situ phenological records as the reference. NDWI gives better estimations of the start of greening up than NDVI (reduced RMSE, bias and dispersions, and higher correlation), whereas it does not improve the determination of the start of leaf coloring. A map of greening up dates in central Siberia obtained from NDWI is shown for year 2002 and the reliability of the method is discussed.     

Remote sensing for detection and monitoring of vegetation affected by oil spills

This study is aimed at demonstrating the application of vegetation spectral techniques for detection and monitoring of the impact of oil spills on vegetation. Vegetation spectral reflectance from Landsat 8 data were used in the calculation of five vegetation indices (normalized difference vegetation index (NDVI), soil adjusted vegetation index (SAVI), adjusted resistant vegetation index 2 (ARVI2), green-infrared index (G-NIR) and green-shortwave infrared (G-SWIR) from the spill sites (SS) and non-spill sites (NSS) in 2013 (pre-oil spill), 2014 (oil spill date) and 2015 (post-oil spill) for statistical comparison. The result shows that NDVI, SAVI, ARVI2, G-NIR and G-SWIR indicated a certain level of significant difference between vegetation condition at the SS and the NSS in December 2013. In December 2014 vegetation conditions indicated higher level of significant difference between the vegetation at the SS and NSS as follows where NDVI, SAVI and ARVI2 with p-value 0.005, G-NIR – p-value 0.01 and G-SWIR p-value 0.05. Similarly, in January 2015 a very significant difference with p-value <0.005. Three indices NDVI, ARVI2 and G-NIR indicated highly significant difference in vegetation conditions with p-value <0.005 between December 2013 and December 2014 at the same sites. Post-spill analysis shows that NDVI and ARVI2 indicated low level of significance difference p-value <0.05 suggesting subtle change in vegetation conditions between December 2014 and January 2015. This technique may help with the real time detection, response and monitoring of oil spills from pipelines for mitigation of pollution at the affected sites in mangrove forests.     

基于主成分分析和植被指数的潮滩植被信息提取研究

针对潮滩湿地植被的特点,利用2004年7月30日的SPOT5数据,在主成分分析的基础上,结合归一化植被指数NDVI和归一化水体指数NDWI,进行了Brovey变换和小波变换融合处理。融合后得到的图像较融合前的图像和原始SPOT5图像在空间相关性、信息熵以及植被间的可分离度等指标上都有显著的提高。对融合效果较佳的基于NDWI和主成分的小波变换的图像进行最大似然法分类,与原始图像最大似然法分类结果相比,分类精度提高了4.41%。     

Disaggregation of LST over India: comparative analysis of different vegetation indices

The non-availability of high-spatial-resolution thermal data from satellites on a consistent basis led to the development of different models for sharpening coarse-spatial-resolution thermal data. Thermal sharpening models that are based on the relationship between land-surface temperature (LST) and a vegetation index (VI) such as the normalized difference vegetation index (NDVI) or fraction vegetation cover (FVC) have gained much attention due to their simplicity, physical basis, and operational capability. However, there are hardly any studies in the literature examining comprehensively various VIs apart from NDVI and FVC, which may be better suited for thermal sharpening over agricultural and natural landscapes. The aim of this study is to compare the relative performance of five different VIs, namely NDVI, FVC, the normalized difference water index (NDWI), soil adjusted vegetation index (SAVI), and modified soil adjusted vegetation index (MSAVI), for thermal sharpening using the DisTrad thermal sharpening model over agricultural and natural landscapes in India. Multi-temporal LST data from Landsat-7 Enhanced Thematic Mapper Plus (ETM+) and Moderate Resolution Imaging Spectroradiometer (MODIS) sensors obtained over two different agro-climatic grids in India were disaggregated from 960 m to 120 m spatial resolution. The sharpened LST was compared with the reference LST estimated from the Landsat data at 120 m spatial resolution. In addition to this, MODIS LST was disaggregated from 960 m to 480 m and compared with ground measurements at five sites in India. It was found that NDVI and FVC performed better only under wet conditions, whereas under drier conditions, the performance of NDWI was superior to other indices and produced accurate results. SAVI and MSAVI always produced poorer results compared with NDVI/FVC and NDWI for wet and dry cases, respectively.     

Foliar nutrient and water content in subtropical tree islands: A new chemohydrodynamic link between satellite vegetation indices and foliar δN values

We examined the relationships between two satellite-derived vegetation indices and foliar δ¹⁵N values obtained from dominant canopy species in a set of tree islands located in the Everglades National Park in South Florida, USA. These tree islands constitute important nutrient hotspots in an otherwise P-limited wetland environment. Foliar δ¹⁵N values obtained from a previous study of 17 tree islands in both slough (perennially wet) and prairie (seasonally wet) locations served as a proxy of P availability at the stand level. We utilized five cloud-free SPOT 4 multispectral images (20 m spatial resolution) from different times of the seasonal cycle to derive two atmospherically corrected vegetation indices: the normalized difference vegetation index (NDVI) and the normalized difference water index (NDWI), averaged for each tree island. NDWI, which incorporates a shortwave infrared (SWIR) band that provides information on leaf water content, showed consistently higher linear fits with island foliar δ¹⁵N values than did NDVI. In addition, NDWI showed greater variation throughout the seasonal cycle than did NDVI, and was significantly correlated with average water stage, which suggests that the SWIR band captures important information on seasonally variable water status. Tree islands in slough locations showed higher NDWI than prairie islands during the dry season, which is consistent with higher levels of transpiration and nutrient harvesting and accumulation for perennially wet locations. Overall, the results suggest that water availability is closely related to P availability in subtropical tree islands, and that NDWI may provide a robust indicator of community-level water and nutrient status.     

Land tessellation effects in mapping agricultural areas by remote sensing at field level

While mapping agricultural areas by remote sensing, it is quite common to operate at cadastral parcel level. Unfortunately, this land tessellation is merely administrative: a single parcel can, in fact, be made of differently managed parts whose spectral properties can be significantly different, being often different their content. In this situation, approaches that aggregate spectral signals of pixels belonging to the same parcel to investigate their average behaviour can generate misleading results. In this work, we evaluated how different field tessellation schemes can condition the interpretation of the spectral behaviour of crops with special concern on time series of NDVI (normalized difference vegetation index) and NDWI (normalized difference water index) spectral indices, which are assumed as proxies of plant vigour and crop/soil water content, respectively. The study relies on Sentinel 2 and Landsat 8 data imaging a rice-cultivated area sited in Piemonte (NW Italy). Two reference land tessellation geometries were taken into account: (a) the local cadastral map (purely administrative land division criterion) and (b) a map obtained by image segmentation of the NDVI time series (purely spectral land division criterion). After signal aggregation, some statistics were therefore computed to test differences both in time (within the same parcel along its temporal profile) and in space (within the same image at different positions at the same time). Results obtained exploring the rice growing season 2016 showed that (a) in 23% (70% at 1 sigma) and 27% (70% at 1 sigma) of segments (respectively for NDVI and NDWI) spectral differences, averagely along the year, are significant, possibly leading to wrong interpretation of occurring dynamics in the area; (b) in rice-cultivated fields, spectral differences suffer from seasonality with a higher incidence in Spring, when rice agronomic phases are more dynamic and, in the meantime, critical for management.     

A new agricultural drought monitoring index combining MODIS NDWI and day–night land surface temperatures: a case study in China

The vegetation health index (VHI) is a widely utilized remote-sensing-based index for monitoring agricultural drought on the regional or global scale. However, the validity of VHI as a drought detection tool relies on the assumption that the normalized difference vegetation index (NDVI) and land-surface temperature (T_s) at a given pixel will vary inversely over time. This assumption may introduce large uncertainties in VHI for drought monitoring over areas with complex landforms, such as China. In order to monitor agricultural drought over the whole of China, a new drought detection index is suggested in this article, termed the vegetation drought index (VDI). VDI is developed from the classical VHI by substituting NDVI and T_s with the normalized difference water index (NDWI) and day–night T_s difference (?T_s), respectively. Terra Moderate Resolution Imaging Spectroradiometer (MODIS) MOD11C3 and MOD13C2 products from 2001 to 2011, monthly precipitation data from 1970 to 2010, and yearly winter wheat yield data from 2000 to 2012 were utilized to evaluate VDI. Results indicated that (1) many areas in China show a positive correlation between NDVI and T_s, especially in the cold season, whereas most areas have a negative correlation between NDWI and ?T_s; (2) VDI has a significant linear correlation with VHI in areas and periods where the NDVI–T_s correlation and NDWI–?T_s correlation are both negative; (3) VDI presents a significant correlation with 3 and 6 month standardized precipitation indices, which is comparable to VHI; and (4) VDI has a significant correlation with normalized crop yield, and is better than VHI. As an example, the extreme drought event over southwestern China from winter 2009 to spring 2010 was successfully explored by VDI. It is concluded that the new index, VDI, has the potential to monitor agricultural drought over the whole of China, including areas and periods where the NDVI–T_s correlation is non-negative.     

NDWI与NDVI指数在区域干旱监测中的比较分析——以2003年江西夏季干旱为例

植被水分指数NDWI是基于短波红外（SWIR）与近红外（NIR）的归一化比值指数。与NDVI相比，它能有效地提取植被冠层的水分含量；在植被冠层受水分胁迫时，NDWI指数能及时地响应，这对于旱情监测具有重要意义。结合2003年夏季MODIS影像数据和地面气象数据，以江西省内一片农田和一片林地为试验区域，分析比较了NDWI与NDVI距平值在短期旱情监测中的有效性。监测结果表明， NDWI对植被冠层水分信息比NDVI更为敏感；在短期干旱监测中，NDWI指数能准确地反映旱情的时空变化。     

Observation of flooding and rice transplanting of paddy rice fields at the site to landscape scales in China using VEGETATION sensor data

A unique physical feature of paddy rice fields is that rice is grown on flooded soil. During the period of flooding and rice transplanting, there is a large proportion of surface water in a land surface consisting of water, vegetation and soils. The VEGETATION (VGT) sensor has four spectral bands that are equivalent to spectral bands of Landsat TM, and its mid-infrared spectral band is very sensitive to soil moisture and plant canopy water content. In this study we evaluated a VGT-derived normalized difference water index (NDWI VGT =(B3-MIR)/ (B3+MIR)) for describing temporal and spatial dynamics of surface moisture. Twenty-seven 10-day composites (VGT- S10) from 1 March to 30 November 1999 were acquired and analysed for a study area (175 km by 165 km) in eastern Jiangsu Province, China, where a winter wheat and paddy rice double cropping system dominates the landscape. We compared the temporal dynamics and spatial patterns of normalized difference vegetation index (NDVI VGT ) and NDWI VGT . The NDWI VGT temporal dynamics were sensitive enough to capture the substantial increases of surface water due to flooding and rice transplanting at paddy rice fields. A land use thematic map for the timing and location of flooding and rice transplanting was generated for the study area. Our results indicate that NDWI and NDVI temporal anomalies may provide a simple and effective tool for detection of flooding and rice transplanting across the landscape.     

Response of spectral vegetation indices to soil moisture in grasslands and shrublands

The relationships between satellite-derived vegetation indices (VIs) and soil moisture are complicated because of the time lag of the vegetation response to soil moisture. In this study, we used a distributed lag regression model to evaluate the lag responses of VIs to soil moisture for grasslands and shrublands at Soil Climate Analysis Network sites in the central and western United States. We examined the relationships between Moderate Resolution Imaging Spectroradiometer (MODIS)-derived VIs and soil moisture measurements. The Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI) showed significant lag responses to soil moisture. The lag length varies from 8 to 56 days for NDVI and from 16 to 56 days for NDWI. However, the lag response of NDVI and NDWI to soil moisture varied among the sites. Our study suggests that the lag effect needs to be taken into consideration when the VIs are used to estimate soil moisture.     

MODIS spectral signals at a flux tower site: Relationships with high-resolution data, and CO2 flux and light use efficiency measurements

In this study, spectral indices were calculated from single date HyMap (3 m; 126 bands), Hyperion (30 m; 242 bands), ASTER (15/30 m; 9 bands), and a time series of MODIS nadir BRDF-adjusted reflectance (NBAR; 1 km, 7 bands) for a study area surrounding the Tumbarumba flux tower site in eastern Australia. The study involved: a) the calculation of a range of physiologically-based vegetation indices from ASTER, HyMap, Hyperion and MOD43B NBAR imagery over the flux tower site; b) comparison across scales between HyMap, Hyperion and MODIS for the normalized difference water index (NDWI) and the Red-Green ratio; c) analysis of relationships between tower-based flux and light use efficiency (LUE) measurements and seasonal and climatic constraints on growth; and d) examination of relationships between fluxes, LUE and time series of NDVI, NDWI and Red-Green ratio. Strong seasonal patterns of variation were observed in NDWI and Red/Green ratio from MODIS NBAR. Correlations between fine (3 and 30 m) and coarse (1 km) scale indices for a small region around the flux tower site were moderately good for Red/Green ratio, but poor for NDWI. Hymap NDWI values for the understorey canopy were much lower than values for the tree canopy. MODIS NDWI was negatively correlated with CO₂ fluxes during warm and cool seasons. The correlation indicated that surface reflectance, affected by a spectrally bright grassland understorey canopy, was decoupled from growth of trees with access to deep soil moisture. The application of physiologically-based indices at earth observation scale requires careful attention to applicability of band configurations, contribution of vegetation components to reflectance signals, mechanistic relationships between biochemical processes and spectral indexes, and incorporation of ancillary information into any analysis.     

Characterizing post‐fire vegetation recovery of California chaparral using TM/ETM+ time‐series data

A time series of normalized difference vegetation index (NDVI) data derived from 11 TM/ETM+ images was used to examine the recovery characteristics of chaparral vegetation in a small watershed near Santa Barbara, California following a fire event in 1985. The NDVI recovery trajectory was compared to a generalized recovery trajectory of leaf area index (LAI) for the same region, which was established using a chronosequence approach and TM/ETM+ data. Post‐fire NDVI recovery trajectories were derived for the entire catchment and for individual vegetation types. Post‐fire NDVI spatial patterns on each image date were compared to the pre‐fire pattern to determine the extent to which the pre‐fire pattern was re‐established, and the rate of this recovery. Results indicated that the post‐fire recovery trajectory for the catchment area average NDVI was similar to the previously established regional LAI trajectory based on a chronosequence approach. The NDVI recovery was disrupted by drought stress and attained pre‐fire levels approximately 10 years after the fire. Individual vegetation types did not exhibit different rates of recovery and the recovery trajectories were only distinguished by the maximum post‐fire NDVI observed after 10 years. The post‐fire NDVI spatial pattern also showed a systematic return to pre‐fire conditions, but exhibited a more substantial disruption due to drought stress than was the case for the average NDVI recovery trajectory.     

Indicators of Northern Eurasia's land‐cover change trends from SPOT‐VEGETATION time‐series analysis 1998–2005

The Boreal and Tundra ecosystems of the mid to high latitudes are sensitive indicators of environmental impacts from both climate change and direct human activities. This study uses inter‐annual and inter‐seasonal SPOT‐VGT mosaics for recent years from 1998 to 2005 covering the entire boreal ecosystems of northern Eurasia. Linear trends could be detected in the NDVI and NDWI time series that differ by season, land‐cover type, and latitude. Significant positive NDVI trends are described for spring and related negative trends for NDWI over the boreal forest zone. They indicate an earlier onset of the vegetation green‐up. Similar vegetation dynamics can be described for autumn. The tundra ecosystems of the northern Eurasia latitudes exhibit trends of negative NDVI and positive NDWI. This may be explained by earlier snowmelt and increasing amounts of surface water from positive temperature anomalies. The non‐ambiguous coarse‐scale indicators require further detailed studies to identify driving factors and amount for positive feedbacks in boreal ecosystems.     

Detection of forest harvest type using multiple dates of Landsat TM imagery

A simple and relatively accurate technique for classifying time-series Landsat Thematic Mapper (TM) imagery to detect levels of forest harvest is the topic of this research. The accuracy of multidate classification of the normalized difference vegetation index (NDVI) and the normalized difference moisture index (NDMI) were compared and the effect of the number of years (1–3, 3–4, 5–6 years) between image acquisition on forest change accuracy was evaluated. When Landsat image acquisitions were only 1–3 years apart, forest clearcuts were detected with producer's accuracy ranging from 79% to 96% using the RGB-NDMI classification method. Partial harvests were detected with lower producer's accuracy (55–80%) accuracy. The accuracy of both clearcut and partial harvests decreased as time between image acquisition increased. In all classification trials, the RGB-NDMI method produced significantly higher accuracies, compared to the RGB-NDVI. These results are interesting because the less common NDMI (using the reflected middle infrared band) outperformed the more popular NDVI. In northern Maine, industrial forest landowners have shifted from clearcutting to partial harvest systems in recent years. The RGB-NDMI change detection classification applied to Landsat TM imagery collected every 2–3 years appears to be a promising technique for monitoring forest harvesting and other disturbances that do not remove the entire overstory canopy.     

Drought and elevation effects on MODIS vegetation indices in northern Arizona ecosystems

Northern Arizona ecosystems are particularly sensitive to plant-available moisture and have experienced a severe drought with considerable impacts on ecosystems from desert shrub and grasslands to pinyon-juniper and conifer forests. Long-term time-series from the Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) are used to monitor recent regional vegetation activity and temporal patterns across various ecosystems. Surface air temperature, solar radiation and precipitation are used to represent meteorological anomalies and to investigate associated impacts on vegetation greenness. Vegetation index anomalies in the northern Arizona ecosystem have a decreasing trend with increasing surface air temperature and decreasing precipitation. MODIS NDVI and EVI anomalies are likely sensitive to the amount of rainfall for northern Arizona ecosystem conditions, whereas inter-annual variability of surface air temperature accounts for MODIS NDVI anomaly variation. The higher elevation area shows the slow vegetation recovery through trend analysis from MODIS vegetation indices for 2000–2011 within the study domain and along elevation.     

Combining ground-based measurements and MODIS-based spectral vegetation indices to track biomass accumulation in post-fire chaparral

Multi-temporal satellite imagery can provide valuable information on the patterns of vegetation growth over large spatial extents and long time periods, but corresponding ground-referenced biomass information is often difficult to acquire, especially at an annual scale. In this study, we test the relationship between annual biomass estimated using shrub growth rings and metrics of seasonal growth derived from Moderate Resolution Imaging Spectroradiometer (MODIS) spectral vegetation indices (SVIs) for a small area of southern California chaparral to evaluate the potential for mapping biomass at larger spatial extents. These SVIs are related to the fraction of photosynthetically active radiation absorbed by the plant canopy, which varies throughout the growing season and is correlated with net primary productivity. The site had most recently burned in 2002, and annual biomass accumulation measurements were available from years 5 to 11 post-fire. We tested the metrics of seasonal growth using six SVIs: normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), soil adjusted vegetation index (SAVI), normalized difference water index (NDWI), normalized difference infrared index 6 (NDII6), and vegetation atmospherically resistant index (VARI). Several of the seasonal growth metrics/SVI combinations exhibit a very strong relationship with annual biomass, and all SVIs show a strong relationship with annual biomass (R² for base value time series metric ranging from 0.45 to 0.89). Although additional research is required to determine which of these metrics and SVIs are the most promising over larger spatial extents, this approach shows potential for mapping early post-fire biomass accumulation in chaparral at regional scales.     

Dynamics of mountain semi-natural grassland meadows inferred from SPOT-VEGETATION and field spectroradiometer data

Permanent semi-natural grassland meadows (lameiros) are characteristic of the mountain rural landscapes in northeast Portugal and represent the main fodder resource for livestock production. Furthermore, these meadows are recognized for their environmental, historical, cultural and visual landscape value. A monitoring study based on remote-sensing data was implemented to understand the impacts of management practices on the lameiros vegetation dynamics and to analyse changes in vegetation dynamics over the period 1998–2008 in response to inter-annual climatic variability. Ten-day SPOT-VEGETATION (VGT) image composites from this period were used to examine the annual temporal profile using the normalized difference vegetation index (NDVI) and their relationship with ground-based observation of vegetation growth and reflectance inferred with a spectroradiometer. Results show that the NDVI profile fits well the characteristic vegetation growth dynamics and associated management practices in the region. For the period from July 2007 to December 2008, the variation in vegetation height explains 46 to 52% of the variation in NDVI derived respectively from spectroradiometer and VGT data. NDVI referring to dates of specific stages of the vegetation dynamics and management practices in lameiros was tested against climatic variables, for the period 1998–2008. More than 57% of the inter-annual variability of the average NDVI during the lameiros development period can be explained by the mean temperature, and 53% of the variability on the date of occurrence of maximum vegetation development (MVD) can be explained by the mean temperature during the spring period. These results support the analysis of lameiros responses to different scenarios of climate and water management and may support the implementation of more efficient farm activities.     

Yield prediction and waterlogging assessment for tea plantation land using satellite image‐based techniques

Motivated by the operational use of remote sensing in various agricultural crop studies, this study evaluates the application and utility of remote sensing‐based techniques in yield prediction and waterlogging assessment of tea plantation land in the Assam State of India. The potential of widely used vegetation indices like NDVI and SR (simple ratio) and the recently proposed TVI has been evaluated for the prediction of green leaf tea yield and made tea yield based on image‐derived leaf area index (LAI), along with weather parameters. It was observed that the yield model based on the TVI showed the highest correlation (R² = 0.83) with green leaf tea yield. The NDVI‐ and SR‐based models suffered non‐responsiveness when the yield approached maximum. The NDVI and SR showed saturation when the LAI exceeded a magnitude of 4. However, the TVI responded well, even when the LAI exceeded 5, and thus has potential use in the estimation of the LAI of dense vegetation such as some crops and forest where it generally exceeds the threshold value of 4.

An attempt was made for the innovative application of TCT and NDWI in the mapping of waterlogging in tea plantation land. The NDWI in conjunction with TCT offered fairly good accuracy (87%) in the delineation of tea areas prone to waterlogging. This observation indicates the potential of NDWI and TCT in mapping waterlogged areas where the soil has considerable vegetation cover.     

Analysis of vegetation recovery surrounding a restored wetland using the normalized difference infrared index (NDII) and normalized difference vegetation index (NDVI)

Watershed restoration efforts seek to rejuvenate vegetation, biological diversity, and land productivity at Cienega San Bernardino, an important wetland in southeastern Arizona and northern Sonora, Mexico. Rock detention and earthen berm structures were built on the Cienega San Bernardino over the course of four decades, beginning in 1984 and continuing to the present. Previous research findings show that restoration supports and even increases vegetation health despite ongoing drought conditions in this arid watershed. However, the extent of restoration impacts is still unknown despite qualitative observations of improvement in surrounding vegetation amount and vigor. We analyzed spatial and temporal trends in vegetation greenness and soil moisture by applying the normalized difference vegetation index (NDVI) and normalized difference infrared index (NDII) to one dry summer season Landsat path/row from 1984 to 2016. The study area was divided into zones and spectral data for each zone was analyzed and compared with precipitation record using statistical measures including linear regression, Mann–Kendall test, and linear correlation. NDVI and NDII performed differently due to the presence of continued grazing and the effects of grazing on canopy cover; NDVI was better able to track changes in vegetation in areas without grazing while NDII was better at tracking changes in areas with continued grazing. Restoration impacts display higher greenness and vegetation water content levels, greater increases in greenness and water content through time, and a decoupling of vegetation greenness and water content from spring precipitation when compared to control sites in nearby tributary and upland areas. Our results confirm the potential of erosion control structures to affect areas up to 5 km downstream of restoration sites over time and to affect 1 km upstream of the sites.     

Remote sensing of spring phenology in boreal regions: A free of snow-effect method using NOAA-AVHRR and SPOT-VGT data (1982-2004)

Measurements of spring phenological dates in boreal regions using NDVI can be affected by snowmelt. This impacts the analysis of interannual variations in phenology and the estimates of annual carbon fluxes. For these two objectives, snowmelt effect must be removed from the phenological detection. We propose a methodology for determining the date of onset of greening in the 1982-2004 period using SPOT-VEGETATION (VGT) and NOAA Advanced Very High Resolution Radiometer (AVHRR) data. From 1998 onwards, the date of onset of greening is taken as the date at which the Normalized Difference Water Index (NDWI), calculated from SPOT-VGT near and short-wave infrared bands, starts increasing. This index decreases with snowmelt but increases with vegetation greening. For the 1982-2001 period, the date of onset of greening is the date at which AVHRR-NDVI equals a pixel specific threshold (PST), determined using the results of the NDWI method in the years common to the two datasets. The methods are validated using in situ measurements of the dates of leaf appearance. RMSE of 6.7 and 7.8 days, respectively, is found using NDWI-VGT and PST-NOAA methodologies, and the difference between the two methodologies in the common years is small. Very importantly, the dates are not biased. The interannual variations of the 23-year spring phenology dataset on the study area in northern Eurasia are analysed. In average over the study area, an advance of 8 days and a delay of 3.6 days are, respectively, found over the periods 1982-1991 and 1993-2004. These results confirm and complete previous studies about the greening trend, remove the uncertainty due to snow, and may improve carbon budget calculations.