Global Vegetation Moisture Index Research Articles

Evaluation of the monitoring capability of various vegetation indices and mainstream satellite band settings for grassland drought

In the context of global climate change and increasing human activities, grassland drought has become increasingly severe and complex. The monitoring of grassland drought is crucial for reducing drought-related losses and ensuring national ecological security. This study used the coupled PROSPECT and SAIL radiative transfer models (PROSAIL) to simulate canopy reflectance, considering factors such as grassland growth stages and varying drought conditions. Our objective was to reveal the spectral response characteristics of grasslands to varying drought conditions and identify sensitive spectral bands suitable for drought monitoring during different grassland growth stages. We aligned commonly available satellite bands from moderate resolution imaging spectroradiometer (MODIS), Sentinel 2, Landsat 8, WorldView 2, and Gaofen 2 (GF 2) with these sensitive bands to assess the capabilities of existing satellite data for drought monitoring. Furthermore, this research evaluated the suitability of 16 commonly used remote sensing vegetation indices for grassland drought monitoring, including Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Ratio Vegetation Index (RVI), Difference Vegetation Index (DVI), Modified Soil Adjusted Vegetation Index (MSAVI), Atmospherically Resistant Vegetation Index (ARVI), Modified Normalized Difference Water Index (MNDWI), Global Vegetation Moisture Index (GVMI), Land Surface Water Index (LSWI), Visible and Shortwave Infrared Drought Index (VSDI), Water index(WI), Moisture Stress Index(MSI), Normalized Difference Water Index(NDWI), Normalized Difference Infrared Index (NDII), Photochemical Reflectance Index (PRI), and Optimized Soil-Adjusted Vegetation Index (OSAVI). The simulation and analysis results revealed: 1) Grassland in different growth stages exhibit similar sensitivities to certain spectral bands, namely those within the ranges of 540 nm–720 nm, 1250 nm–1690 nm, 1805 nm–2190 nm, and 2264 nm–2500 nm, which are more sensitive to various drought conditions. 2) Suitable vegetation indices for both growing and stable stages include NDII, MSI, PRI, LSWI, and GVMI, with silhouette coefficients exceeding 0.6 for the growing stage and 0.7 for the stable stage. The least suitable vegetation index is DVI, with an average silhouette coefficient of 0.15 over the entire growth stage. 3) From the spectral band perspective, among the five assessed satellites, MODIS Band 7 exhibits the highest sensitivity to water content across all satellite bands. MODIS's band configuration is most suitable for monitoring grassland drought during different growth stages, while WorldView 2's band configuration is the least suitable.

A Method for Estimating Alfalfa (Medicago sativa L.) Forage Yield Based on Remote Sensing Data

Alfalfa (Medicago sativa L.) is a widely planted perennial legume forage plant with excellent quality and high yield. In production, it is very important to determine alfalfa growth dynamics and forage yield in a timely and accurate manner. This study focused on inverse algorithms for predicting alfalfa forage yield in large-scale alfalfa production. We carried out forage yield and aboveground biomass (AGB) field surveys at different times in 2022. The correlations among the reflectance of different satellite remote sensing bands, vegetation indices, and alfalfa forage yield/AGB were analyzed, additionally the suitable bands and vegetation indices for alfalfa forage yield inversion algorithms were screened, and the performance of the statistical models and machine learning (ML) algorithms for alfalfa forage yield inversion were comparatively analyzed. The results showed that (1) regarding different harvest times, the alfalfa forage yield inversion model for first-harvest alfalfa had relatively large differences in growth, and the simulation accuracy of the alfalfa forage yield inversion model was higher than that for the other harvest times, with the growth of the second- and third-harvest alfalfa being more homogeneous and the simulation accuracy of the forage yield inversion model being relatively low. (2) In the alfalfa forage yield inversion model based on a single parameter, the moisture-related vegetation indices, such as the global vegetation moisture index (GVMI), normalized difference water index (NDWI) and normalized difference infrared index (NDII), had higher coefficients of correlation with alfalfa forage yield/AGB, and the coefficients of correlation R2 values for the first-harvest alfalfa were greater than 0.50, with the NDWI correlation being the best with an R2 value of 0.60. (3) For the alfalfa forage yield inversion model constructed with vegetation indices and band reflectance as multiparameter variables, the random forest (RF) and support vector machine (SVM) simulation accuracy was higher than that of the alfalfa forage yield inversion model based on a single parameter; the first-harvest alfalfa R2 values based on the multiparameter RF and SVM models were both 0.65, the root mean square errors (RMSEs) were 329.74 g/m2 and 332.32 g/m2, and the biases were −0.47 g/m2 and −2.24 g/m2, respectively. The vegetation indices related to plant water content can be considered using a single parameter inversion model for alfalfa forage yield, the vegetation indices and band reflectance can be considered using a multiparameter inversion model for alfalfa forage yield, and ML algorithms are also an optimal choice. The findings in this study can provide technical support for the effective and strategic production management of large-scale alfalfa.

The role of drought conditions on the recent increase in wildfire occurrence in the high Andean regions of Peru

Wildfire occurrence has increased sharply in the last two decades in the Peruvian Andes. There is, however, little research on wildfires and their impacts. This study explores the conditions conducive to wildfire during 2020. MODIS images were collected to estimate the development of vegetation. In addition, ground-based monthly and satellite-based daily precipitation data were collected. Daily precipitation regularity was evaluated using a concentration index (CI), while monthly precipitation was used to estimate the Standard Precipitation Index (SPI). We used also the Global Vegetation Moisture Index (GVMI), which is a useful indicator of vegetation dynamics based on vegetation moisture. Our results do not indicate a direct link between rainfall regularity (lowest CI values) and development of vegetation. Although the SPI drought analysis using seasonal rainfall indicated nearly normal conditions during 2019–2020, analysis of dry-day frequency (DDF) suggests that the dry period played an important role between September and November 2020, producing conditions similar to the droughts of 2005, 2010 and 2016. GVMI also showed below-average values from April to November. We corroborate the usefulness of DDF for monitoring the potential increase in wildfire conditions. A controlled burn policy could offer a more useful way to reduce the impacts of wildfire.

Remote sensing-based assessment of ecosystem health by optimizing vigor-organization-resilience model: A case study in Fuzhou City, China

Although remote sensing technology has been used to evaluate regional ecosystem health for a long time, it is still necessary to find a suitable index system to better evaluate ecosystem health. This study aims to improve the ecosystem health measurement ability of remote sensing technology. This research was carried out in Fuzhou under the traditional Vigor-Organization-Resilience (VOR) framework by optimizing and improving the construction method of sub-indexes. Sub-indexes were constructed using spectral index analysis, landscape theoretical ecology model and spatial measurement. Three remote sensing datasets were used (1996, 2008 and 2021) to carry out remote sensing diagnosis of regional ecosystem health in the Fuzhou administrative region. The main research findings and conclusions were as follows. A new comprehensive vigor index (CVI) was developed by the principal component analysis (PCA) based on the four indicators: fractional vegetation cover (FVC), global vegetation moisture index (GVMI), vegetation temperature condition index (VTCI), normalized differential build-up and bare soil index (NDBSI). A new organizational index was constructed based on the landscape index. Four types of indexes, namely landscape heterogeneity (LH), landscape connectivity (LC), the shape characteristics of forest patches (CS) and the connectivity of forest patches (CC) were used as the main factors for calculating the organizational index. A resilience index calculation framework was proposed based on the habitat quality model. The temporal and spatial characteristics of ecosystem health were evaluated and analyzed. The regional ecosystem health value of the whole region reduced gradually, with average values of 0.3521 (1996), 0.3445 (2008) and 0.3345 (2021) respectively. The average reduction rate was 0.0007 per year (1996–2021). The proposed remote sensing diagnosis method provides a complete framework for solving the problems of measuring the dynamic evolution process and characteristics of regional ecosystem health.

Spatial pattern prediction of forest wildfire susceptibility in Central Yunnan Province, China based on multivariate data

Wildfires are an important disturbance factor in forest ecosystems. Assessing the probability of forest wildfires can assist in forest wildfire prevention, control, and supervision. The logistic regression model is widely used to forecast the probability, spatial patterns, and drivers of forest wildfires. This study used logistic regression to establish a spatial prediction model for forest wildfire susceptibility, which was applied to evaluate the risk of forest wildfires in Central Yunnan Province (CYP), China. A forest wildfire risk classification was implemented for CYP using forest burn scar data for 2001 to 2020 and the logistic spatial prediction model for forest wildfire susceptibility. Climate, vegetation, topographical, human activities, and location were selected as forest wildfire prediction variables. The results showed that: (1) The distributions of temperature, vegetation coverage, distance to water bodies, distance to roads, and precipitation were positively correlated with the occurrence of forest wildfires. Elevation, relative humidity, the global vegetation moisture index, wind speed, slope, latitude, and distance to residential areas were negatively correlated with the occurrence of forest wildfires. (2) The results of the logistic spatial prediction model for forest wildfire susceptibility showed a good fit to wildfire data, with an overall simulation probability of 81.6%. The optimal threshold for spatial prediction for forest wildfire susceptibility in CYP was determined to be 0.414. A significance level of a selected model variable of < 0.05 resulted in an area under the receiver operating characteristic curve (AUC) of 0.882–0.890. (3) Forest wildfire prevention efforts should focus on Southwest Yuxi City and southern Qujing City accounted for a high proportion of the areas at high risk of forest wildfires. Other localities should adjust forest wildfire prevention measures according to local conditions and strengthen existing wildfire prevention and emergency resource planning and allocation. (4) Some factors contributing to forest wildfires where different among the different areas. Forest wildfire risk factors had different degrees of impact under different spatial and temporal scales. The spatial relationships between wildfire disasters and influencing factors should be established in areas with heterogeneous environmental conditions for the selection of relevant factors.

On the interchangeability of Landsat and MODIS data in the CMRSET actual evapotranspiration model – Comment on “Monitoring irrigation using Landsat observations and climate data over regional scales in the Murray-Darling Basin” by David Bretreger, In-Young Yeo, Greg Hancock and Garry Willgoose

In a recent paper published in this journal, Bretreger et al. (2020) estimate irrigation water use from satellite remotely sensed estimates of actual evapotranspiration in five irrigated districts of the Murray-Darling Basin (southeast Australia). They used three models that scale crop reference evapotranspiration with vegetation indices acquired by recent Landsat satellites: (i) IrriSAT, (ii) Kamble and (iii) CMRSET. In their paper, irrigation water use computed with CMRSET generally overestimated observed irrigation water use, sometimes fivefold. Based on these results, Bretreger et al. (2020) discouraged the use of CMRSET for irrigation monitoring. In this comment, we reproduce the experiments in Bretreger et al. (2020), and demonstrate that their overestimation was because they used an incorrect Landsat band in their implementation of CMRSET. CMRSET was originally calibrated using both MODIS-derived Enhanced Vegetation Index (EVI) and Global Vegetation Moisture Index (GVMI). To calculate GVMI, a shortwave infrared (SWIR) band with a wavelength of ∼1.6 μm was used, which for MODIS is named SWIR2, and for Landsat is named SWIR1 (i.e., different sensors have different bands and a different number of bands and SWIR1 in MODIS has a wavelength of ∼1.2 μm). In their Landsat CMRSET implementation, Bretreger et al. (2020) computed GVMI using the Landsat SWIR2 band, which has a wavelength of ∼2.1 μm (see their Table 3). We show that CMRSET implemented with the Landsat SWIR1 band (i.e., with a wavelength of ∼1.6 μm, so being the correct Landsat band to calculate GVMI) yields similar results (both for temporal patterns and magnitude) when compared to the other two remote sensing actual evapotranspiration models in Bretreger et al. (2020). For the irrigation districts, these similar results meant that the mean water year (i.e., July to next June from 2010 to 2017) actual evapotranspiration mean absolute relative difference was 15.3% (with a 5.4%–26.5% range) and the water year irrigation water use mean absolute relative difference was 4.1% (with a 9.8%–18.9% range). Conversely, we show that a Landsat CMRSET implementation using the incorrect Landsat band for the calculation of GVMI (i.e., with a wavelength of ∼2.1 μm) led to a large overestimation agreeing with the CMRSET results reported by Bretreger et al. (2020), both for water year actual evapotranspiration (mean absolute relative difference of 64.6%, with a 51.1%–83.6% range) and irrigation water use (mean absolute relative difference of 44.1%, with a 31.4%–65.1% range). The use of the incorrect Landsat band means that the Bretreger et al. (2020) recommendation specific to their Landsat CMRSET implementation is invalid. A demonstration of Bretreger et al.’s (2020) incorrect Landsat CMRSET implementation and the correct one can be accessed at: https://jorgepena.users.earthengine.app/view/comment-on-bretreger-2020-joh-paper.Despite Bretreger et al.’s (2020) shortcomings in their implementation of the Landsat CMRSET actual evapotranspiration model, timely monitoring of irrigation water use via satellite remote sensing is required at Landsat resolutions (i.e., 30 m) or higher due to the heterogeneous nature of irrigation in most agricultural landscapes, and we applaud them for pursuing this line of research.

Hurricane Irma impact on biophysical and biochemical features of canopy vegetation in the Santa Fe River Basin, Florida

Changes of biophysical and biochemical features in canopy cover over space and time can be used to elucidate watershed resilience to climate change or extreme weather. Geospatial technologies such as satellite remote sensing are helpful to identify such vegetation dynamics. This paper investigates the landfall impact of Hurricane Irma on canopy cover in the Santa Fe River Basin (Florida), which occurred in August 2017. Specifically, this study explores the effect of Hurricane Irma on vegetation dynamics via biophysical and biochemical features during the pre and post hurricane landfall. The geospatial analysis compares a suite of remote sensing spectral indices including enhanced vegetation index (EVI), leaf area index (LAI), fraction of photosynthetically active radiation (FPAR), evapotranspiration (ET), land surface temperature (LST), and global vegetation moisture index (GVMI). It is noticeable that paired spectral indices such as EVI-LST, GVMI-LST, LAI-LST, EVI-GVMI, EVI-LAI, LAI-FPAR, EVI-FPAR, and LAI-GVMI reflecting the dependent relationships between biophysical and biochemical features reveal varying level of logistic trends under Irma landfall impact. In addition, the evolution of these biophysical and biochemical features associated with the vegetation cover was analyzed in terms of the functional capacity over grassland, evergreen forested land, deciduous forested land, and agricultural land, to understand watershed response following Irma landfall. The functional capacity of the ecosystem reduced in terms of EVI and LAI over these four land-over patterns due to the landfall of Hurricane Irma. The interactions between agricultural land variation and flooding impact before and after the hurricane landfall further entail human impact on natural system in a food-water nexus.

The Green Revolution from space: Mapping the historic dynamics of main rice types in one of the world's food bowls

This paper develops a methodology to map the two main rice types in northwest Bangladesh from 1989 to 2016, when Green Revolution technologies and policies resulted in a 300% rice area expansion and localised unsustainable groundwater use. The mapping is performed for the largely irrigated dry season Boro rice (grown from November/December to May/June), and the largely rainfed Aman rice (grown from June/July to October/November).The petabyte archive of Landsat reflectance data available via Google Earth Engine is used to extract monthly time-series of two remotely sensed vegetation indices – the Enhanced Vegetation Index (EVI) and the Global Vegetation Moisture Index (GVMI) – which can capture crop phenological dynamics as they correlate during crop growth stages and can be used to discriminate standing water when EVI is low and GVMI is high. This feature is exploited to detect the effects of flooding before transplanting Boro rice in January and to capture the late phenological phases of Aman rice through harvesting months (October to November).A gap-filled and smoothed EVI and GVMI monthly time-series sequences for both rice cropping months are grouped using machine learning unsupervised K-means clustering into clusters (25 used here), and manually aggregated using crop calendars and local expert knowledge into rice types, other vegetated areas, water, non-permanent water and bare areas. A representative subset of these data are then used to construct a training sample using additional covariates besides the monthly EVI and GVMI values and train a Random Forest (RF) model capable of predicting the land cover types in other years.The resulting RF maps for the two cropping seasons were evaluated against publicly available survey crop statistics (from 1989 to 2016) and published Boro 30 m resolution maps for 2013 to 2016. The maps provide qualitative evidence that unsustainable groundwater extractions can be linked to Boro rice irrigation in some parts of northwest Bangladesh. Further research, using these maps and other hydrological data, is required to quantitatively assess when, where, and how much irrigation contributes to unsustainable groundwater use. The results show that the methodology can accurately (generally within 20% absolute error) capture Boro and Aman rice over the Green Revolution period, and other landscape dynamic characteristics of importance for environmental assessments, such as areas of standing water.

Potential conditions for fire occurrence in vegetation in the Peruvian Andes

Fire activity in the Peruvian Andes has increased significantly in recent decades, but climatic parameters associated with drought, which may indirectly contribute to the occurrence of severe forest fires, have not yet been investigated. Because fire prevention tools are scarce, strategies for deterring burning are necessary in order to reduce impacts in regions where forest fires usually result from human activity. This study explores the conditions conducive to forest fire in the Andes of Peru. Daily precipitation and temperature observed data from the PISCO gridded dataset for the 2002–2016 period were used. In addition, MODIS satellite images (MOD09A1 product) were collected to characterise Andean vegetation using spectral indices. Analysis of daily temperature and rainfall indicates that climatic parameters such as cumulative precipitation, dry-day frequency and hot-day frequency are statistically associated with conditions that could contribute to increased forest fire occurrence. Our findings suggest that a decrease in the water content of vegetation, estimated by the Global Vegetation Moisture Index during the dry period and wet period onset, can be used to identify potential conditions for forest fire occurrence. This study suggests that forest managers should consider implementing prevention strategies that include continuous monitoring of climate and vegetation parameters.

The Daily Fire Hazard Index: A Fire Danger Rating Method for Mediterranean Areas

Mediterranean forests are gravely affected by wildfires, and despite the increased prevention effort of competent authorities in the past few decades, the yearly number of fires and the consequent damage has not decreased significantly. To this end, a number of dynamical methods have been developed in order to produce short-term hazard indices, such as the Fire Probability Index and the Fire Weather Index. The possibility to estimate the fire hazard is based on the observation that there is a relationship between the characteristics of the vegetation (i.e., the fuel), in terms of abundance and moisture content, and the probability of fire insurgence. The density, type, and moisture content of the vegetation are modeled using custom fuel maps, developed using the latest Corine Land Cover, and using a number of indices such as the NDVI (Normalized Difference Vegetation Index), Global Vegetation Moisture Index (GVMI), and the evapotranspiration, derived from daily satellite imagery. This paper shows how the algorithm for the calculation of the Fire Potential Index (FPI) was improved by taking into account the effect of wind speed, topography, and local solar illumination through a simple temperature correction, preserving the straightforward structure of the FPI algorithm. The results were validated on the Italian region of Sardinia using official wildfire records provided by the regional administration.

Sugarcane drought detection through spectral indices derived modeling by remote-sensing techniques

Several indices based on satellite images have been explored to monitor agricultural drought. Despite the existence of some drought indices, no drought monitoring system for sugarcane exists. In this sense, drought detection could be useful tool to quantify losses and help with action plans. This study investigates the Landsat image potential for sugarcane drought detection by assessing the relationship between vegetation and agricultural drought indices (normalized difference vegetation index (NDVI), vegetation condition index (VCI), normalized difference water index (NDWI), global vegetation moisture index (GVMI), and normalized difference infrared index (NDII)). Two new indices combining near-infrared (NIR) and short-wave infrared (SWIR) bands are proposed for sugarcane drought detection. All indices were individually and collectively compared with soil water deficit and water surplus, simulated by the climatological soil–water balance (CSWB) model. A significant correlation between spectral indices and water balance results, specifically for NDVI and VCI indices (~ 30%), was observed. The drought detection system identification was developed by cluster analysis classifying the pixels into three distinct groups (drought, intermediate drought, and non-drought) to later be used in the discriminant analysis. This methodology showed to have an accuracy rate of 65%. However, the discriminant analysis approach was better suited for sugarcane drought monitoring when compared with individual spectral indices.

Assessing the role of SWIR band in detecting agricultural crop stress: a case study of Raichur district, Karnataka, India.

The potential consequence of global climatic change caused the rise in temperature and precipitation decline, the aftermath of which has led to phenomenon like drought. As the agriculture is mainly dependent on the timely availability of water, delayed arrival of rainfall or decrease in the intensity of precipitation highly affects the growth of crops. If such situation persists for a longer period, the soil moisture content will be exploited completely and maturity of the crop will be stunted, finally adversely affecting the annual crop yield. In the present study, the capability of shortwave-infrared (SWIR) channels in detecting the agricultural crop stress in Raichur district of Karnataka state, India, using spectral vegetation indices namely Global Vegetation Moisture Index (GVMI) and Normalized Multi-band Drought Index (NMDI) has been analyzed using multi-temporal MODIS data. The vegetation health analysis by utilizing both indices were carried out from year 2002 to 2012 for the Kharif season, and it was seen that the year 2002 suffered major agricultural drought where the lower GVMI and NMDI values were covering the majority of the crop areas and in the year 2010 agricultural crop production was observed to be good. The average values of GVMI and NMDI for the years 2002 and 2010 were plotted with the average NDVI values of all months of both years and the results revealed that NDVI values were in concurrence with both NMDI and GVMI.

Estimation of daily vegetation coefficients using MODIS data for clear and cloudy sky conditions

ABSTRACTSpatially distributed vegetation coefficients () data with high temporal resolution are in demand for actual evapotranspiration estimation, crop condition assessment, irrigation scheduling, etc. Traditional remotely sensed based data application gets hindered because of two main reasons i.e 1) spectral reflectance based accounts only for transpiration factor, but fails to account for total evapotranspitration. 2) required optical spectral reflectances are available only during clear sky conditions, which creates gaps in the data. Hence there is a necessity of a model which accounts for both transpiration and evpaoration factors and also for a gap filling method, which can produce accurate continuous quantification of values. Therefore, in this study, different combinations of enhanced vegetation index (EVI), global vegetation moisture index (GVMI) and temperature vegetation dryness index (TVDI) have been employed in linear and non linear regression techniques to obtain the best model. To fill the gaps in the data, initially, temporal fitting of values have been examined using Savitsky-Goley (SG) filter for 3 years of data (2012–2014), but this fails when sufficient high quality values are unavailable. In this regard, three gap filling techniques namely regression, artificial neural networks (ANN) and interpolation techniques have been employed over Cauvery basin. Microwave polarization difference index (MPDI) has been employed in ANN technique to estimate values under cloudy sky conditions. The results revealed that the combination of GVMI and TVDI using linear regression technique performed better than other combinations with correlation coefficient (r) and root mean square error (RMSE) values of 0.824 and 0.204 respectively. Furthermore, the results indicated that SG filter can be used for temporal fitting and for gap filling regression technique performed better than other techniques with the r and RMSE values of 0.68 and 0.25 for Berambadi station.

Evaluation of microwave remote sensing for monitoring live fuel moisture content in the Mediterranean region

Live fuel moisture content (LFMC) is an important factor in fire risk management in the Mediterranean region. Drawing upon a large network of stations (the Réseau Hydrique) measuring LFMC for operational fire danger assessment in the south-eastern region of France, this study assesses the ability of several long-term passive microwave remote sensing indices to capture the LFMC temporal dynamic of various Mediterranean shrub species. Microwave remote sensing has a high potential for monitoring LFMC independently of several constraints (e.g., atmospheric and cloud contamination effects) associated with optical-infrared and thermal remote sensing observations. The following four microwave-derived indices are considered: (1) the Essential Climate Variable near-surface soil moisture (ECV_SM); (2) the root-zone soil moisture (ECV_RZSM) derived from ECV_SM; (3) the microwave polarization difference index (MPDI) computed from five microwave frequencies (C, X, Ku, K and Ka-band corresponding to 6.9, 10.7, 18.7, 23.8 and 36.5GHz respectively); and (4) the vegetation optical depth (VOD) at C- and X-band (from the Advanced Microwave Scanning Radiometer for the Earth observing system, AMSR-E). Firstly, an evaluation of the root-zone soil moisture ECV_RZSM against a network of soil moisture measurements (SMOSMANIA in southern France) gave satisfactory results. For most of the Réseau Hydrique sites, the present study found good agreement between LFMC and individual microwave indices, including root-zone soil moisture, VOD at X-band, and MPDI at X and Ku-bands, all averaged over the 15days preceding the in-situ LFMC measurements. VOD at X-band showed the best agreement with the in situ LFMC data (median of correlation coefficients over all in situ sites=0.43). Further comparisons between LFMC data and several optical indices computed from the Moderate Resolution Imaging Spectrometer (MODIS) data including normalized difference vegetation index (NDVI), soil adjusted vegetation index (SAVI), visible atmospheric resistant index (VARI), normalized difference water index (NDWI), normalized difference infrared index 6 (NDII6), normalized difference infrared index 7 (NDII7) and global vegetation moisture index (GVMI) were made. The comparisons showed that VARI and SAVI, as optical greenness indices, outperform the microwave indices and other optical indices with median of correlation coefficients of 0.66 and 0.65, respectively. Overall, this study shows that passive microwave indices, particularly VOD, are efficient proxies for LFMC of Mediterranean shrub species and could be used along with optical indices to evaluate fire risks in the Mediterranean region.

Building probabilistic models of fire occurrence and fire risk zoning using logistic regression in Shanxi Province, China

Fires are a recurrent environmental and economic emergency throughout the world. Fire risk analysis and forest fire risk zoning are important aspects of forest fire management. MODIS remote sensing datasets for Shanxi Province from 2002 to 2012 were used to build a spatial logistic forest fire risk model, based on the spatial distribution of forest fires and forest fire-influencing factors, using geographic information system technology. A forest fire risk zoning study was conducted at a large temporal scale and a provincial spatial scale. The resulting logistic model of forest fire risk, built with spatial sampling, showed a good fit (p < 0.05) between the distribution of forest fires and forest fire impact factors. The relative operating characteristic value was 0.757, and a probability distribution map for forest fire was developed, using layer computing. The forest fire area of Shanxi Province was divided into zones of zero, low, moderate, high and extremely high fire risk. The influences of altitude (GC), land-use type (LT), land surface temperature (LST), normalized difference vegetation index (NDVI) and global vegetation moisture index (GVMI) on fire events presented significant spatial variability, whereas the influences of slope and distance to the nearest path exhibited insignificant spatial variability in Shanxi Province. The influences of NDVI and LST on fire events were significant throughout Shanxi Province, whereas the influences of GC, LT and GVMI were only significant locally. Seven fire-prevention regions were delineated, based on the fire-influencing factors. Different fire-prevention policies and emphases should be taken into consideration for each of the seven fire-prone regions.

Evaluation of wildfire propagation susceptibility in grasslands using burned areas and multivariate logistic regression

This research simulated wildfire propagation susceptibility based on multivariate logistic regression. Moderate Resolution Imaging Spectrometer (MODIS)-derived fuel indicators and topographic factors were the independent variables, and burnt areas served as the dependent variable. MODIS data were collected daily during the wildfire seasons of April to May and September to October from 2001 to 2007 to acquire information about live and dead fuel in the Mongolia–China grasslands. The inputs for the independent parameters for wildfire propagation susceptibility modelling were the normalized difference vegetation index (NDVI), optimized soil-adjusted vegetation index (OSAVI), moisture stress index (MSI), global vegetation moisture index (GVMI), dead fuel INDEX (DFI), elevation, slope, and aspect. Multivariate logistic regression ranking indicates that DFI, MSI, DEM, and OSAVI are the top four factors, with an overall accuracy of 80%. ‘Leave one out’ cross-validation demonstrated that the overall accuracy of the propagation susceptibility modelling ranged from 65% to 87%. Finally, the model was used to produce 10 day average wildfire propagation susceptibility maps during the wildfire seasons of 2001–2007 and to predict the location of burned areas. This research will be useful for understanding the propagation susceptibility of wildfires in grassland areas and for creating policies for preventing wildfire spread.

Estimation of Herbaceous Fuel Moisture Content Using Vegetation Indices and Land Surface Temperature from MODIS Data

The monitoring of herbaceous fuel moisture content is a crucial activity in order to assess savanna fire risks. Faced with the difficulty of managing wide areas of vegetated surfaces, remote sensing appears an attractive alternative for terrestrial measurements because of its advantages related to temporal resolution and spatial coverage. Earth observation (EO)-based vegetation indices (VIs) and the ratio between Normalized Difference Vegetation Index (NDVI) and surface temperature (ST) were used for assessment of herbaceous fuel moisture content estimates and validated against herbaceous data collected in 2010 at three open savanna sites located in Senegal, West Africa. EO-based estimates of water content were more consistent with the use of VI as compared to the ratio NDVI/ST. Different VIs based on near-infrared (NIR) and shortwave infrared (SWIR) reflectance were tested and a consistent relationship was found between field measurements of leaf equivalent water thickness (EWT) from all test sites and Normalized Difference Infrared Index (NDII), Global Vegetation Moisture Index (GVMI) and Moisture Stress Index (MSI). Also, strong relationships were found between fuel moisture content (FMC) and VIs for the sites separately; however, they were weaker for the pooled data. The correlations between EWT/FMC and VIs were found to decrease progressively as the woody cover increased. Although these results suggest that NIR and SWIR reflectance can be used for the estimation of herbaceous water content, additional validation from an increased number of study sites is necessary to study the robustness of such indices for a larger variety of savanna vegetation types.

Developing a gross primary production model for coniferous forests of northeastern USA from MODIS data

Accurate estimation of ecosystem carbon fluxes is crucial for understanding the feedbacks between the terrestrial biosphere and the atmosphere and for making climate-policy decisions. A statistical model is developed to estimate the gross primary production (GPP) of coniferous forests of northeastern USA using remotely sensed (RS) radiation (land surface temperature and near-infra red albedo) and ecosystem variables (enhanced vegetation index and global vegetation moisture index) acquired by the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor. This GPP model (called R-GPP-Coni), based only on remotely sensed data, was first calibrated with GPP estimates derived from the eddy covariance flux tower of the Howland forest main tower site and then successfully transferred and validated at three other coniferous sites: the Howland forest west tower site, Duke pine forest and North Carolina loblolly pine site, which demonstrate its transferability to other coniferous ecoregions of northeastern USA. The proposed model captured the seasonal dynamics of the observed 8-day GPP successfully by explaining 84–94% of the observed variations with a root mean squared error (RMSE) ranging from 1.10 to 1.64gC/m2/day over the 4 study sites and outperformed the primary RS-based GPP algorithm of MODIS.

Modeling gross primary production of deciduous forest using remotely sensed radiation and ecosystem variables

We explored the potential application of two remotely sensed (RS) variables, the Global Vegetation Moisture Index (GVMI) and the near‐infrared albedo (AlbedoNIR), in modeling the gross primary production (GPP) of three deciduous forests. For the Harvard Forest (deciduous) of Massachusetts, it was found that GPP is strongly correlated with GVMI (coefficient of determination, R2 = 0.60) during the growing season, and with AlbedoNIR (R2 = 0.82) throughout the year. Subsequently, a statistical model called the Remotely Sensed GPP (R‐GPP) model was developed to estimate GPP using remotely sensed radiation (land surface temperature (LST), AlbedoNIR) and ecosystem variables (enhanced vegetation index (EVI) and GVMI). The R‐GPP model, calibrated and validated against the GPP estimates derived from the eddy covariance flux tower of the Harvard Forest, could explain 95% and 92% of the observed GPP variability for the study site during the calibration (2000–2003) and the validation (2004–2005) periods, respectively. It outperformed the primary RS‐based GPP algorithm of Moderate Resolution Imaging Spectroradiometer (MODIS), which explained 80% and 77% of the GPP variability during 2000–2003 and 2004–2005, respectively. The calibrated R‐GPP model also explained 93% and 94% of the observed GPP variation for two other independent validation sites, the Morgan Monroe State Forest and the University of Michigan Biological Station, respectively, which demonstrates its transferability to other deciduous ecoregions of northeastern United States.

Scaling of potential evapotranspiration with MODIS data reproduces flux observations and catchment water balance observations across Australia

We developed a new algorithm for estimating monthly actual evapotranspiration (AET) based on surface reflectance from MODIS-Terra and interpolated climate data. The algorithm uses monthly values of the Enhanced Vegetation Index (EVI) and the Global Vegetation Moisture Index (GVMI) derived from the MODIS nadir bidirectional reflectance distribution function – adjusted reflectance product (MOD43B4) to scale Priestley-Taylor potential evapotranspiration derived from the climate surfaces. The EVI is associated with evapotranspiration through its relationship with leaf area index. The GVMI allows separation between surface water and bare soil when EVI is low and provides information on vegetation water content when EVI is high. The model was calibrated using observed AET data from seven sites in Australia, including two forests, two open savannas, a grassland, a floodplain and a lake. Model outputs were compared with four year average difference between precipitation and streamflow (a surrogate for mean AET) in 227 unimpaired catchments across Australia. We tested four different model configurations and found that the best results both in the calibration and evaluation datasets were obtained when a precipitation interception term ( E i ) and the GVMI were incorporated into the model. The E i term and the GVMI improved AET estimates in the forest, savanna and grassland sites and in the lake and floodplain sites respectively. The most comprehensive model estimated monthly AET at the seven calibration sites with a RMSE of 18.0 mm mo −1 (22% of the mean AET, r 2 = 0.84). In the evaluation dataset, mean annual AET was estimated with a RMSE of 137.44 mm y −1 (19% of the mean AET, r 2 = 0.61). The model was able to reproduce the main spatial and temporal patterns in AET across Australia. The main advantages of the proposed model are that it uses a single set of parameters (i.e. does not need an auxiliary land cover map) and that it is able to estimate AET in areas with significant direct evaporation, including lakes and floodplains.