Analysis of multi-temporal SPOT NDVI images for small-scale land-use mapping

Bie, C.A.J.M. de; Khan, Mobushir Riaz; Smakhtin, Vladimir; Venus, V.; Weir, M.J.C.; Smaling, E.M.A.

doi:10.1080/01431161.2010.512939

Cited by 51 publications

(43 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other studies have discussed and demonstrated that using hyper-temporal normalized difference vegetation index (NDVI) image datasets is particularly suitable to map (agro-) ecosystems, exploiting their ability to capture aspects of vegetation phenology and crop calendar characteristics (Gorham 1998, Murakami et al 2001, Uchida 2001, Hill and Donald 2003, de Bie 2004, de Bie et al 2008, 2011, Sakamoto et al 2009, Khan et al 2010.…”

Section: Hyper-temporal Ndvi Imagesmentioning

confidence: 98%

LaHMa: a landscape heterogeneity mapping method using hyper-temporal datasets

Bie

Nguyen

Ali

et al. 2012

International Journal of Geographical Information Science

View full text Add to dashboard Cite

A new quantitative method extracts a landscape heterogeneity map (LaHMa) from hyper-temporal remote-sensing data. The feature extraction method is data-driven, unbiased, and builds on the commonly used data reduction technique of Iterative Self-Organizing Data Analysis (ISODATA) clustering with the support of divergence separability indices. First, the relevant spatial-temporal variation in normalized difference vegetation index (NDVI) is classified through ISODATA clustering. Second, a series of prepared cluster maps are overlaid to examine and detect the frequency with which boundaries between clusters occur at the same location. This step identifies the boundary strength between clusters and detects spatial heterogeneity within them. Results of the method are explored for the typical agriculture-defined landscape of the Mekong delta, Vietnam, using NDVI-imagery time-series from SPOT-Vegetation and MODIS-Terra. The method extracts useful landscape heterogeneity features and can support land-cover mapping requiring information on fragmentation and land-cover gradients.

show abstract

Section: Hyper-temporal Ndvi Imagesmentioning

confidence: 98%

LaHMa: a landscape heterogeneity mapping method using hyper-temporal datasets

Bie

Nguyen

Ali

et al. 2012

International Journal of Geographical Information Science

View full text Add to dashboard Cite

show abstract

“…In China, an increasing number of studies have been carried out to map vegetation types using Pan et al 2003;Wang & Tenhunen 2004;Xiao et al 2005;Jiang et al 2008). Several recent studies have confirmed that SPOT-VGT NDVI and moderate-resolution imaging spectroradiometer (MODIS) NDVI imagery can be used to produce the required land-cover and vegetation maps at various scales (Bagan et al 2007;Matsuoka et al 2007;Fu et al 2010;de Bie et al 2011;Nguyen et al 2012). In this study, we aimed to map at macro-habitat scale the fraction of evergreen forest and the presence probability of epiphyllous liverworts using time-series SPOT-VGT NDVI imagery, and to assess the quantitative relationship between the fraction of evergreen forest and the presence probability of epiphyllous liverworts.…”

Section: Introductionmentioning

confidence: 99%

Hyper‐temporal remote sensing helps in relating epiphyllous liverworts and evergreen forests

Jiang

Bie

Wang

et al. 2012

J Vegetation Science

View full text Add to dashboard Cite

Question Is there, at the macro‐habitat scale, a relationship between the fraction of evergreen forests and the presence probability of epiphyllous liverworts? Can these two parameters be estimated and mapped using an NDVI‐based indicator that is derived from time‐series of SPOT‐VGT imagery? Location Southern China. Methods Applying the ISODATA algorithm, we classified SPOT‐VGT NDVI time‐series imagery and produced an NDVI map at 1‐km2 resolution containing 128 NDVI classes. That map and the National Land Cover Map of China (2000) were used to prepare a scheme for field sampling. For 537 1‐km2 areas, located in 19 blocks of 100 km × 100 km, field data were collected. They represented 51 preselected NDVI classes that were assumed to contain evergreen forests. Data on the fractions of forest and evergreen forest, the fraction of evergreen forest in forest present and presence probability of epiphyllous liverworts were regressed against different NDVI class‐specific indicators by means of weighted least‐squares regression (WLSR). Results The SPOT‐VGT NDVI for March was found to best explain the variation between 1‐km2 areas in the presence probability of epiphyllous liverworts (R2 = 0.64; linear relationship; RMSE = 0.38) as the area fraction (%) of evergreen forest (R2 = 0.90; exponential relationship; RMSE = 6.1%). Epiphyllous liverworts were only found within 1‐km2 areas when the SPOT‐VGT NDVI value for March was more than 0.50 (model estimate: 0.43 ± 0.20) and the fraction of evergreen forest in 1 km2 was above 14% (model estimate: 5 ± 35%). The estimation errors (with 95% confidence interval) of these two relationships were calculated using a bootstrap resampling method with 1000 replications; they were, respectively, 0.49–0.76 and 0.85–0.94 for R2, and 0.11–0.23 and 5.2–11.9% for RMSE. Other positive relationships were found between the presence probability of epiphyllous liverworts and the fraction of evergreen forest (R2 = 0.64, linear relationship; RMSE = 0.38) and between the fraction of evergreen forest and forest within 1‐km2 areas (R2 = 0.80, linear relationship; RMSE = 29%). Conclusion We show that for southern China, the fraction of evergreen forest and the presence probability of epiphyllous liverworts can directly be inferred by making use of SPOT‐VGT NDVI imagery. The findings are fully consistent with earlier reported hotspots for epiphyllous liverworts. At the macro‐habitat scale, the presence probability of epiphyllous liverworts proved quantitatively related to the fraction of evergreen forest. This suggests causal relationships with patch size and/or rainfall patterns. We believe that the derived maps may serve as a foundation for a range of further studies.

show abstract

“…Fourth, we generated a stratification layer based on the unsupervised classification of NDVI images from eMODIS (see Section 2.2.2), applying an unsupervised clustering algorithm, namely the Iterative Self-Organizing Data Analysis Technique (ISODATA), similar to the approach of de Bie et al [43]. The aim was to obtain a biophysical stratification purely relying on RS data.…”

Section: Linear Regression Model and Spatial Aggregation Levelsmentioning

confidence: 99%

Phenology-Based Biomass Estimation to Support Rangeland Management in Semi-Arid Environments

et al. 2017

View full text Add to dashboard Cite

Livestock plays an important economic role in Niger, especially in the semi-arid regions, while being highly vulnerable as a result of the large inter-annual variability of precipitation and, hence, rangeland production. This study aims to support effective rangeland management by developing an approach for mapping rangeland biomass production. The observed spatiotemporal variability of biomass production is utilised to build a model based on ground and remote sensing data for the period 2001 to 2015. Once established, the model can also be used to estimate herbaceous biomass for the current year at the end of the season without the need for new ground data. The phenology-based seasonal cumulative Normalised Difference Vegetation Index (cNDVI), computed from 10-day image composites of the Moderate-resolution Imaging Spectroradiometer (MODIS) NDVI data, was used as proxy for biomass production. A linear regression model was fitted with multi-annual field measurements of herbaceous biomass at the end of the growing season. In addition to a general model utilising all available sites for calibration, different aggregation schemes (i.e., grouping of sites into calibration units) of the study area with a varying number of calibration units and different biophysical meaning were tested. The sampling sites belonging to a specific calibration unit of a selected scheme were aggregated to compute the regression. The different aggregation schemes were evaluated with respect to their predictive power. The results gathered at the different aggregation levels were subjected to cross-validation (cv), applying a jackknife technique (leaving out one year at a time). In general, the model performance increased with increasing model parameterization, indicating the importance of additional unobserved and spatially heterogeneous agro-ecological effects (which might relate to grazing, species composition, optical soil properties, etc.) in modifying the relationship between cNDVI and herbaceous biomass at the end of the season. The biophysical aggregation scheme, the calibration units for which were derived from an unsupervised ISODATA classification utilising 10-day NDVI images taken between January 2001 and December 2015, showed the best performance in respect to the predictive power (R 2 cv = 0.47) and the cross-validated root-mean-square error (398 kg·ha −1 ) values, although it was not the model with the highest number of calibration units. The proposed approach can be applied for the timely production of maps of estimated biomass at the end of the growing season before field measurements are made available. These maps can be used for the improved management of rangeland resources, for decisions on fire prevention and aid allocation, and for the planning of more in-depth field missions.

show abstract

Analysis of multi-temporal SPOT NDVI images for small-scale land-use mapping

Cited by 51 publications

References 28 publications

LaHMa: a landscape heterogeneity mapping method using hyper-temporal datasets

LaHMa: a landscape heterogeneity mapping method using hyper-temporal datasets

Hyper‐temporal remote sensing helps in relating epiphyllous liverworts and evergreen forests

Phenology-Based Biomass Estimation to Support Rangeland Management in Semi-Arid Environments

Contact Info

Product

Resources

About