Estimating crop area using seasonal time series of enhanced vegetation index from MODIS satellite imagery

Cereal grain is one of the main export commodities of Australian agriculture. Over the past decade, crop yield forecasts for wheat and sorghum have shown appreciable utility for industry planning at shire, state, and national scales. There is now an increasing drive from industry for more accurate and cost-effective crop production forecasts. In order to generate production estimates, accurate crop area estimates are needed by the end of the cropping season. Multivariate methods for analysing remotely sensed Enhanced Vegetation Index (EVI) from 16-day Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery within the cropping period (i.e. April-November) were investigated to estimate crop area for wheat, barley, chickpea, and total winter cropped area for a case study region in NE Australia. Each pixel classification method was trained on ground truth data collected from the study region. Three approaches to pixel classification were examined: (i) cluster analysis of trajectories of EVI values from consecutive multi-date imagery during the crop growth period; (ii) harmonic analysis of the time series (HANTS) of the EVI values; and (iii) principal component analysis (PCA) of the time series of EVI values. Images classified using these three approaches were compared with each other, and with a classification based on the single MODIS image taken at peak EVI. Imagery for the 2003 and 2004 seasons was used to assess the ability of the methods to determine wheat, barley, chickpea, and total cropped area estimates. The accuracy at pixel scale was determined by the percent correct classification metric by contrasting all pixel scale samples with independent pixel observations. At a shire level, aggregated total crop area estimates were compared with surveyed estimates. All multi-temporal methods showed significant overall capability to estimate total winter crop area. There was high accuracy at pixel scale (>98% correct classification) for identifying overall winter cropping. However, discrimination among crops was less accurate. Although the use of single-date EVI data produced high accuracy for estimates of wheat area at shire scale, the result contradicted the poor pixel-scale accuracy associated with this approach, due to fortuitous compensating errors. Further studies are needed to extrapolate the multi-temporal approaches to other geographical areas and to improve the lead time for deriving cropped-area estimates before harvest.

Measuring and predicting canopy nitrogen nutrition in wheat using a spectral index—The canopy chlorophyll content index (CCCI).

Varying the spatial distribution of applied nitrogen (N) fertilizer to match demand in crops has been shown to increase profits in Australia. Better matching the timing of N inputs to plant requirements has been shown to improve nitrogen use efficiency and crop yields and could reduce nitrous oxide emissions from broad acre grains. Farmers in the wheat production area of south eastern Australia are increasingly splitting N application with the second timing applied at stem elongation (Zadoks 30). Spectral indices have shown the ability to detect crop canopy N status but a robust method using a consistent calibration that functions across seasons has been lacking. One spectral index, the canopy chlorophyll content index (CCCI) designed to detect canopy N using three wavebands along the "red edge" of the spectrum was combined with the canopy nitrogen index (CNI), which was developed to normalize for crop biomass and correct for the N dilution effect of crop canopies. The CCCI-CNI index approach was applied to a 3-year study to develop a single calibration derived from a wheat crop sown in research plots near Horsham, Victoria, Australia. The index was able to predict canopy N (g m-2) from Zadoks 14-37 with an r2 of 0.97 and RMSE of 0.65 g N m-2 when dry weight biomass by area was also considered. We suggest that measures of N estimated from remote methods use N per unit area as the metric and that reference directly to canopy %N is not an appropriate method for estimating plant concentration without first accounting for the N dilution effect. This approach provides a link to crop development rather than creating a purely numerical relationship. The sole biophysical input, biomass, is challenging to quantify robustly via spectral methods. Combining remote sensing with crop modelling could provide a robust method for estimating biomass and therefore a method to estimate canopy N remotely. Future research will explore this and the use of active and passive sensor technologies for use in precision farming for targeted N management.

Invasion impacts on biodiversity: responses of ant communities to infestation by cat’s claw creeper, Macfadyena unguis-cati (Bignoniaceae) in subtropical Australia

Ants are the dominant soil faunal group in many if not most terrestrial ecosystems, and play a key role in soil structure and function. This study documents the impacts of invasion by the exotic cat’s claw creeper vine, Macfadyena unguis-cati (L.) Gentry (Bignoniaceae) on surface-situated (epigaeic) and subterranean (hypogaeic) ant communities in subtropical SE Queensland Australia where it is a major environmental weed of riparian areas, rainforest communities and remnant natural vegetation, smothering standing vegetation and causing canopy collapse. Soil ants were sampled in infested and uninfested areas at eight sites spanning both riparian and non-riparian habitats in subtropical SE Queensland. Patterns of ant species composition and functional grouping in response to patch invasion status, landscape type and habitat stratum were investigated using ANOVA and non-metric multidimensional scaling ordination. The epigaeic and subterranean strata supported markedly different ant assemblages, and ant communities also differed between riparian and non-riparian habitats. However, M. unguis-cati invasion had a surprisingly limited impact. There was a tendency for ant abundance and species richness to be lower in infested patches, and overall species composition was different between infested and uninfested patches, but these differences were relatively small, and did not occur consistently across sites. There were changes in functional group composition that conformed to known functional group responses to environmental change, but these were similarly limited and inconsistent across sites. Our study has shown that ant communities are surprisingly resilient to invasion by M. unguis-cati, and serves as a warning against making assumptions about invasion impacts based on visual appearances.

Estimating winter crop area across seasons and regions using time-sequential MODIS imagery.

The wheat grain industry is Australia's second largest agricultural export commodity. There is an increasing demand for accurate, objective and near real-time crop production information by industry. The advent of the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite platform has augmented the capability of satellite-based applications to capture reflectance over large areas at acceptable pixel scale, cost and accuracy. The use of multi-temporal MODIS-enhanced vegetation index (EVI) imagery to determine crop area was investigated in this article. Here the rigour of the harmonic analysis of time-series (HANTS) and early-season metric approaches was assessed when extrapolating over the entire Queensland (QLD) cropping region for the 2005 and 2006 seasons. Early-season crop area estimates, at least 4 months before harvest, produced high accuracy at pixel and regional scales with percent errors of -8.6% and -26% for the 2005 and 2006 seasons, respectively. In discriminating among crops at pixel and regional scale, the HANTS approach showed high accuracy. The errors for specific area estimates for wheat, barley and chickpea were 9.9%, -5.2% and 10.9% (for 2005) and -2.8%, -78% and 64% (for 2006), respectively. Area estimates of total winter crop, wheat, barley and chickpea resulted in coefficient of determination (R(2)) values of 0.92, 0.89, 0.82 and 0.52, when contrasted against the actual shire-scale data. A significantly high coefficient of determination (0.87) was achieved for total winter crop area estimates in August across all shires for the 2006 season. Furthermore, the HANTS approach showed high accuracy in discriminating cropping area from non-cropping area and highlighted the need for accurate and up-to-date land use maps. The extrapolability of these approaches to determine total and specific winter crop area estimates, well before flowering, showed good utility across larger areas and seasons. Hence, it is envisaged that this technology might be transferable to different regions across Australia.

Light thresholds derived from seagrass loss in the coastal zone of the northern Great Barrier Reef, Australia

There is a world-wide trend for deteriorating water quality and light levels in the coastal zone, and this has been linked to declines in seagrass abundance. Localized management of seagrass meadow health requires that water quality guidelines for meeting seagrass growth requirements are available. Tropical seagrass meadows are diverse and can be highly dynamic and we have used this dynamism to identify light thresholds in multi-specific meadows dominated by Halodule uninervis in the northern Great Barrier Reef, Australia. Seagrass cover was measured at similar to 3 month intervals from 2008 to 2011 at three sites: Magnetic Island (MI) Dunk Island (DI) and Green Island (GI). Photosynthetically active radiation was continuously measured within the seagrass canopy, and three light metrics were derived. Complete seagrass loss occurred at MI and DI and at these sites changes in seagrass cover were correlated with the three light metrics. Mean daily irradiance (I-d) above 5 and 8.4 mol m(-2) d(-1) was associated with gains in seagrass at MI and DI, however a significant correlation (R = 0.649, p < 0.05) only occurred at MI. The second metric, percent of days below 3 mol m(-2) d(-1), correlated the most strongly (MI, R = -0.714, p < 0.01 and DI, R = -0.859, p = <0.001) with change in seagrass cover with 16-18% of days below 3 mol m(-2) d(-1) being associated with more than 50% seagrass loss. The third metric, the number of hours of light saturated irradiance (H-sat) was calculated using literature-derived data on saturating irradiance (E-k). H-sat correlated well (R = 0.686, p <0.01; and DI, R = 0.704, p < 0.05) with change in seagrass abundance, and was very consistent between the two sites as 4 H-sat was associated with increases in seagrass abundance at both sites, and less than 4 H-sat with more than 50% loss. At the third site (GI), small seasonal losses of seagrass quickly recovered during the growth season and the light metrics did not correlate (p > 0.05) with change in percent cover, except for I-d which was always high, but correlated with change in seagrass cover. Although distinct light thresholds were observed, the departure from threshold values was also important. For example, light levels that are well below the thresholds resulted in more severe loss of seagrass than those just below the threshold. Environmental managers aiming to achieve optimal seagrass growth conditions can use these threshold light metrics as guidelines; however, other environmental conditions, including seasonally varying temperature and nutrient availability, will influence seagrass responses above and below these thresholds. (C) 2012 Published by Elsevier Ltd.

Projecting future expansion of invasive species: Comparing and improving methodologies for species distribution modeling

Modeling the distributions of species, especially of invasive species in non-native ranges, involves multiple challenges. Here, we developed some novel approaches to species distribution modeling aimed at reducing the influences of such challenges and improving the realism of projections. We estimated species-environment relationships with four modeling methods run with multiple scenarios of (1) sources of occurrences and geographically isolated background ranges for absences, (2) approaches to drawing background (absence) points, and (3) alternate sets of predictor variables. We further tested various quantitative metrics of model evaluation against biological insight. Model projections were very sensitive to the choice of training dataset. Model accuracy was much improved by using a global dataset for model training, rather than restricting data input to the species’ native range. AUC score was a poor metric for model evaluation and, if used alone, was not a useful criterion for assessing model performance. Projections away from the sampled space (i.e. into areas of potential future invasion) were very different depending on the modeling methods used, raising questions about the reliability of ensemble projections. Generalized linear models gave very unrealistic projections far away from the training region. Models that efficiently fit the dominant pattern, but exclude highly local patterns in the dataset and capture interactions as they appear in data (e.g. boosted regression trees), improved generalization of the models. Biological knowledge of the species and its distribution was important in refining choices about the best set of projections. A post-hoc test conducted on a new Partenium dataset from Nepal validated excellent predictive performance of our “best” model. We showed that vast stretches of currently uninvaded geographic areas on multiple continents harbor highly suitable habitats for Parthenium hysterophorus L. (Asteraceae; parthenium). However, discrepancies between model predictions and parthenium invasion in Australia indicate successful management for this globally significant weed. This article is protected by copyright. All rights reserved.

Projecting future expansion of invasive species: Comparing and improving methodologies for species distribution modeling

Modeling the distributions of species, especially of invasive species in non-native ranges, involves multiple challenges. Here, we developed some novel approaches to species distribution modeling aimed at reducing the influences of such challenges and improving the realism of projections. We estimated species-environment relationships with four modeling methods run with multiple scenarios of (1) sources of occurrences and geographically isolated background ranges for absences, (2) approaches to drawing background (absence) points, and (3) alternate sets of predictor variables. We further tested various quantitative metrics of model evaluation against biological insight. Model projections were very sensitive to the choice of training dataset. Model accuracy was much improved by using a global dataset for model training, rather than restricting data input to the species’ native range. AUC score was a poor metric for model evaluation and, if used alone, was not a useful criterion for assessing model performance. Projections away from the sampled space (i.e. into areas of potential future invasion) were very different depending on the modeling methods used, raising questions about the reliability of ensemble projections. Generalized linear models gave very unrealistic projections far away from the training region. Models that efficiently fit the dominant pattern, but exclude highly local patterns in the dataset and capture interactions as they appear in data (e.g. boosted regression trees), improved generalization of the models. Biological knowledge of the species and its distribution was important in refining choices about the best set of projections. A post-hoc test conducted on a new Partenium dataset from Nepal validated excellent predictive performance of our “best” model. We showed that vast stretches of currently uninvaded geographic areas on multiple continents harbor highly suitable habitats for Parthenium hysterophorus L. (Asteraceae; parthenium). However, discrepancies between model predictions and parthenium invasion in Australia indicate successful management for this globally significant weed. This article is protected by copyright. All rights reserved.