G. E. Donald

Estimating spatio-temporal patterns of agricultural productivity in fragmented landscapes using AVHRR NDVI time series

Abstract The characteristics of Normalized Difference Vegetation Index (NDVI) time series can be disaggregated into a set of quantitative metrics that may be used to derive information about vegetation phenology and land cover. In this paper, we examine the patterns observed in metrics calculated for a time series of 8 years over the southwest of Western Australia—an important crop and animal production area of Australia. Four analytical approaches were used; calculation of temporal mean and standard deviation layers for selected metrics showing significant spatial variability; classification based on temporal and spatial patterns of key NDVI metrics; metrics were analyzed for eight areas typical of climatic and production systems across the agricultural zone; and relationships between total production and productivity measured by dry sheep equivalents were developed with time integrated NDVI (TINDVI). Two metrics showed clear spatial patterns; the season duration based on the smooth curve produced seven zones based on increasing length of growing season; and TINDVI provided a set of classes characterized by differences in overall magnitude of response, and differences in response in particular years. Frequency histograms of TINDVI could be grouped on the basis of a simple shape classification: tall and narrow with high, medium or low mean indicating most land is responsive agricultural cover with uniform seasonal conditions; broad and short indicating that land is of mixed cover type or seasonal conditions are not spatially uniform. TINDVI showed a relationship to agricultural productivity that is dependent on the extent to which crop or total agricultural production was directly reduced by rainfall deficiency. TINDVI proved most sensitive to crop productivity for Statistical Local Areas (SLAs) having rainfall less than 600 mm, and in years when rainfall and crop production were highly correlated. It is concluded that metrics from standardized NDVI time series could be routinely and transparently used for retrospective assessment of seasonal conditions and changes in vegetation responses and cover.

Integration of optical and radar classifications for mapping pasture type in Western Australia

In this study, independent classifications of Landsat Thematic Mapper imagery and Jet Propulsion Laboratory AirSAR were combined to create an integrated classification of pasture and other vegetation types for a study area in the agricultural zone of Western Australia. The resulting classification combines greenness and brightness information from optical data with structure and water content information from synthetic aperture radar (SAR). Field observations of vegetation type, botanical composition, ground cover percentage, wet and dry biomass, canopy height, and soil water content were collected at 34 sites representing a range of pastures, browse shrubs, and crops. An unsupervised version of the Complex Wishart classification procedure, based on preserving scattering characteristics from the Freeman and Durden backscatter decomposition, was applied to the C-, L-, and P-band polarimetric SAR data. The optical classification was carried out using a principle component analysis on the green, red, and near-infrared bands and clustering on the basis of a class centroid distance measure and knowledge of ground targets. These two classification results were then fused together. Assessment of a confusion matrix using the individual sites showed that identification of more uniform, dense, and structurally distinct canopies was better than that of more diverse, sparse, and structurally ambiguous canopies, as the former were better represented by the canopy height attribute used in the SAR classification component. The optical classification enabled correction of SAR misclassification of vegetation due to surface roughness and soil moisture effects, or similar backscatter responses from herbaceous or arboreal canopies. The results show that simplification of vegetation into groups based upon properties with sensitive responses in both the optical and SAR domains, and combination of separate SAR and optical classifications, has potential for improving classification of diverse and heterogeneous herbaceous and browse cover in grazing lands. However, collection of ground calibration data must be at an appropriate spatial scale and include canopy and surface measurements directly related to backscatter mechanisms and spectral sensitivity.

Pasture Land Cover in Eastern Australia from NOAA-AVHRR NDVI and Classified Landsat TM

Abstract A method for constructing a pasture land cover classification for the high rainfall zone of eastern Australia from the U.S. National Oceanic and Atmospheric Administration’s advanced very high resolution radiometer normalized difference vegetation index (NOAA-AVHRR NDVI) data is described. The method uses an established classification of pasture growth potential from single date, springtime, Landsat thematic mapper (TM) data to provide measures of subpixel mixture composition in mosaicked classes in the absence of ground truth or control sites. A sequence of AVHRR data from 1993 was transformed into a vegetation index and then classified to define pasture classes with differing patterns of NDVI. High-resolution classifications of pasture were constructed for ten selected sites within the study area by using Landsat TM scenes. The study area was split into a northern and southern zone on the basis of the temporal pattern of moisture indices. The pasture land cover classes were described in terms of the shape of the NDVI profiles, their geographical location, and the subpixel composition from Landsat TM data. The classified NDVI data were combined with local government area (LGA) boundary data to allow the particular pasture state of each LGA to be estimated. The NDVI–Landsat TM procedure identified 21 and 22 classes in the northern and southern zones, respectively. These classes could be broadly grouped into eight types: sown perennial pastures, sown perennial pastures with woodland, sown annual pastures (southern zone only), mixed pasture and cropping, native pastures, native pastures with woodland, degraded or revegetated areas, and forest. This eight-class classification combining the two zones appeared to represent regional distribution of the major types quite well. The pasture land cover classification was evaluated for selected LGAs by using agricultural statistics and a specialist pasture survey. Local estimates of proportions of major pasture types were sometimes inaccurate owing to difficulties in distinguishing between perennial, annual, and native types where seasonal conditions caused rapid senescence or where open woodland confused profiles between improved and native pastures. The method is nevertheless useful where ground truth or definitive spectral signatures for cover types are unavailable and where description in terms of an average or predominant cover type within a landscape mosaic is acceptable.

Evaluating an active optical sensor for quantifying and mapping green herbage mass and growth in a perennial grass pasture

Efficiently measuring and mapping green herbage mass using remote sensing devices offers substantial potential benefits for improved management of grazed pastures over space and time. Several techniques and instruments have been developed for estimating herbage mass, however, they face similar limitations in terms of their ability to distinguish green and senescent material and their use over large areas. In this study we explore the application of an active, near infrared and red reflectance sensor to quantify and map pasture herbage mass using a range of derived spectral indices. The Soil Adjusted Vegetation Index offered the best correlation with green dry matter (GDM), with a root mean square error of prediction of 288 kg/ha. The calibrated sensor was integrated with a Global Positioning System on a 4-wheel motor bike to map green herbage mass. An evaluation of representative, truncated transects indicated the potential to conduct rapid assessments of the GDM in a paddock, without the need for full paddock surveys.

Relating Radar Backscatter to Biophysical Properties of Temperate Perennial Grassland

Abstract The response of polarimetric airborne synthetic aperture radar (SAR) to grassland is investigated. Synthetic aperture radar from the National Aeronautical and Space Administration/ Jet Propulsion Laboratory (NASA/JPL) airborne imaging system was acquired over a diverse grassland site in northern NSW, Australia in September 1993. Grassland and high backscatter targets are classified using images from C, L, and P band for hh, hv, and vv polarizations. The grassland classes cover a wide dynamic range of backscatter from −7 dB to −14 dB in C band and −9 dB to −23 dB in L band. Significant regression relationships are formulated between measurements of grassland height and radar backscatter using site data aggregated by 25 mm height class. The relationship between species composition and grassland classes is explored. Polarization effects include an enhanced range of backscatter across grassland classes for 45° cross polarization at L band and differences in the pedestal height of the C band polarization signature for species classes. The results of drainage modeling suggest that soil moisture is a significant confounding factor influencing radar backscatter from the grassland. Simple models using logistic probability of association between height class and radar backscatter in a Bayesian inference engine, and a simple threshold based on logistic probability of association between wet areas and P vv are examined. Our results suggest that combined imagery from C and L band satellite-borne SAR sensors have potential for current application in grassland monitoring.

Quantitative mapping of pasture biomass using satellite imagery

A knowledge of the amount of pasture biomass available in farm paddocks is crucial for improving utilization and productivity in the Australian grazing industry. A method to quantitatively map the biomass of annual pastures under grazing has been developed using the Normalized Difference Vegetation Index (NDVI) derived from high-resolution satellite imagery. Relationships between field-measured pasture biomass and the NDVI were examined for different transects in paddocks under different grazing regimes across three geographically dispersed farm sites. A significant linear relationship (R 2 = 0.84) was observed when the NDVI was regressed against biomass. The slope of the relationship between the NDVI and biomass declined in a highly predictable (R 2 = 0.82) exponential form as the growing season progressed and this pattern was consistent across four separate seasons. This knowledge was used to formulate a reliable model to predict paddock average pasture biomass using the NDVI. The model estimates were validated against observed biomass in the range 500–4000 kilograms of dry matter per hectare (kg DM ha–1) with R 2 = 0.85 and a standard error of 315 (kg DM ha–1).

Using MODIS imagery, climate and soil data to estimate pasture growth rates on farms in the south-west of Western Australia.

Remote sensing of vegetation and its monitoring using the normalised difference vegetation index (NDVI) offers the opportunity to provide a coverage of agricultural land at a large scale. The availability of MODIS NDVI at a resolution of 250 m provided the opportunity to evaluate the hypothesis that pasture growth rate (PGR) of individual paddocks can be accurately predicted using a model based on MODIS NDVI in combination with climate and soil data and a light-use efficiency model. Model estimates of PGR were compared with field measurements of PGR recorded in grazing enclosure cages collected over 3 years from six farms located across the south-west region of Western Australia. The estimates attained from the model explained 70% of the variation in PGR for individual paddocks on farms over the 3 years of the study, with an average error at the paddock scale of 10.4 kg DM/ha.day over all growing seasons and years. Across all farms studied, there was generally good agreement between satellite-derived PGR and ground-based measurements, although estimates of PGR varied between years and farms. The model explained 47% of the variation in pasture growth early in the season (from break of season to end of July), compared with 62% late in the season (from August to pasture senescence). The present study demonstrated that PGR for individual paddocks can be predicted at weekly intervals from MODIS imagery, climate and soil data and a light-use efficiency model at an accuracy sufficient to facilitate on-farm pasture and livestock management.

Near real-time Feed On Offer (FOO) from MODIS for early season grazing management of Mediterranean annual pastures

Near real-time estimation of Feed On Offer (FOO) from Moderate Resolution Imaging Spectroradiometer (MODIS) data was developed to help farmers improve their grazing management during early growth of annual pastures to maximize grass utilization for wool production. Data were collected from 72 fields on 15 farms in southwestern Australia. From these data, an exponential relationship at the field scale between the Normalized Difference Vegetation Index (NDVI) estimated from MODIS and FOO data was derived for the vegetative growth phase for FOO between 0 and 2000 kg ha–1 (R 2 = 0.71–0.75). This relationship transformed the dimensionless index NDVI to a dimensioned (kg ha–1) measure of FOO, from which farmers could apply extension advice received from the Western Australian (WA) Department of Agriculture and Food Production (DAFP). Above an FOO of 2000 kg ha–1 or when the annual pasture species began to senesce, the relationship ceased to have predictive value. Near real-time estimates of FOO from MODIS proved useful to farmers despite an apparent standard error of ±300 kg ha–1. How to reduce the errors in FOO predicted from MODIS NDVI is also discussed.

The spectral response of pastures in an intensively managed dairy system

All grazing-based industries require information on their feed resources in order to manage them optimally. Gathering this information through traditional methods for measuring pasture biomass is time-consuming and error-prone, resulting in increased interest in remotely-sensed methods. Remote sensing used to monitor feed resources in farming systems differs from remote sensing of systems such as forestry because of how the time-scale of management practices impacts on the growth rate and accumulation patterns of biomass. Also, in operational systems, designed for near real-time delivery to end-users of quantitative pasture measurements, we are restricted to the commercially available broad-band high-resolution sensors. The goal of this paper is to understand how remotely-sensed observations of pastures in an intensively managed dairy system change in relation to intensive management practices, so that better image analysis and ground-validation methods can be developed for measuring and monitoring such systems. At two dates in the growing season we examined high-resolution (SPOT-5 and Ikonos) images of an intensively managed perennial dairy farm in Victoria (Australia). We showed that the observed spectral response in the images varied with the length of time since the paddock was grazed, consistent with the re-growth of pastures post-grazing. The operational remote sensing of pastures is often restricted by the range of spectral bands that are available on broad-band sensors. However, these results suggest that when choosing a vegetation index for intensively managed dairy pastures it should incorporate the short-wave infrared (SWIR) band to improve observations of recently grazed pastures and tune analyses based on the spectral response.

Ground truthing protocols for biomass estimation in rangeland environments

Remote sensing for the assessment and mapping of total standing biomass relies on accurate ground data for calibration and validation. The spatial heterogeneity of rangelands pose challenges in sampling methodologies, demanding a large number of replicate measurements that are expensive and labour demanding when working on the scale of pastoral stations. In this paper we present a ground truthing protocol that can be used for biomass estimation in heterogeneous rangeland environments, important for the development of assessments based on remote sensing or growth modelling. The protocol is based on a combination of visual estimates, crop circle NDVI, and disk-plate meter height recordings. Relationships between these indirect measurements and biomass were specific for either season or vegetation type. A combination of these measurements in a multivariate regression provided an accurate alternative, while strongly reducing the number of cuts required.