Inferential, non-parametric statistics to assess quality of probabilistic forecast systems

Many statistical forecast systems are available to interested users. In order to be useful for decision-making, these systems must be based on evidence of underlying mechanisms. Once causal connections between the mechanism and their statistical manifestation have been firmly established, the forecasts must also provide some quantitative evidence of `quality’. However, the quality of statistical climate forecast systems (forecast quality) is an ill-defined and frequently misunderstood property. Often, providers and users of such forecast systems are unclear about what ‘quality’ entails and how to measure it, leading to confusion and misinformation. Here we present a generic framework to quantify aspects of forecast quality using an inferential approach to calculate nominal significance levels (p-values) that can be obtained either by directly applying non-parametric statistical tests such as Kruskal-Wallis (KW) or Kolmogorov-Smirnov (KS) or by using Monte-Carlo methods (in the case of forecast skill scores). Once converted to p-values, these forecast quality measures provide a means to objectively evaluate and compare temporal and spatial patterns of forecast quality across datasets and forecast systems. Our analysis demonstrates the importance of providing p-values rather than adopting some arbitrarily chosen significance levels such as p < 0.05 or p < 0.01, which is still common practice. This is illustrated by applying non-parametric tests (such as KW and KS) and skill scoring methods (LEPS and RPSS) to the 5-phase Southern Oscillation Index classification system using historical rainfall data from Australia, The Republic of South Africa and India. The selection of quality measures is solely based on their common use and does not constitute endorsement. We found that non-parametric statistical tests can be adequate proxies for skill measures such as LEPS or RPSS. The framework can be implemented anywhere, regardless of dataset, forecast system or quality measure. Eventually such inferential evidence should be complimented by descriptive statistical methods in order to fully assist in operational risk management.

Development and Accreditation of an Applied Climate Education Unit for Sustainable Land Use in Australia

In recent years, there have been significant developments in climate science relevant to agriculture and natural resource management. Assessing impacts of climate variability and use of seasonal climate forecasts have become increasingly important elements in the management "toolkit" for many Australian farmers. Consideration of climate change further increases the need for improved management strategies. While climate risk extension activities have kept pace with advances in climate science, a national review of the Vocational Education and Training system in Australia in relation to "weather and climate" showed that these topics were "poorly represented" at the management level in the Australian Qualifications Framework, and needed increased emphasis. Consequently, a new Unit of Competency concerning management of climatic risk was developed and accredited to address this deficiency. The objective of the unit was to build knowledge and skills for better management of climate variability via the elements of surveying climatic and enterprise data; analysing climatic risks and opportunities; and developing climatic risk management strategies. This paper describes establishment of a new unit for vocational education that is designed to harness recent developments in applied climate science for better management of Australia's highly variable climate. The main benefits of the new unit of competency, "Developing climatic risk management strategies,"were seen as improving decisions in climate and agriculture, and reducing climate risk exposure to enhance sustainable agriculture. The educational unit is now within the scope of agricultural colleges, universities, and registered training organisations as an accredited unit.

An evaluation of genetic analyses, skull morphology and visual appearance for assessing dingo purity: implications for dingo conservation

The introgression of domestic dog genes into dingo populations threatens the genetic integrity of 'pure' dingoes. However, dingo conservation efforts are hampered by difficulties in distinguishing between dingoes and hybrids in the field. This study evaluates consistency in the status of hybridisation (i.e. dingo, hybrid or dog) assigned by genetic analyses, skull morphology and visual assessments. Of the 56 south-east Queensland animals sampled, 39 (69.6%) were assigned the same status by all three methods, 10 (17.9%) by genetic and skull methods, four (7.1%) by genetic and visual methods; and two (3.6%) by skull and visual methods. Pair-wise comparisons identified a significant relationship between genetic and skull methods, but not between either of these and visual methods. Results from surveying 13 experienced wild dog managers showed that hybrids were more easily identified by visual characters than were dingoes. A more reliable visual assessment can be developed through determining the relationship between (1) genetics and phenotype by sampling wild dog populations and (2) the expression of visual characteristics from different proportions and breeds of domestic dog genes by breeding trials. Culling obvious hybrids based on visual characteristics, such as sable and patchy coat colours, should slow the process of hybridisation.

Encroachment of Echinococcus granulosus into urban areas in eastern Queensland, Australia

Objective: To investigate the prevalence of Echinococcus granulosus in wild dogs (dingos and dingo-domestic dog hybrids) living in and around human habitation on Fraser Island and in townships of the Maroochy Shire, on Queensland's Sunshine Coast, Australia. Design: Wild dogs were humanely killed on Fraser Island and in the Maroochy Shire because they were deemed a potential danger to the public. Their intestines were collected and the contents examined for intestinal parasites. Procedure: Intestines were removed as soon after death as possible, packed in plastic bags and kept frozen until examination. The intestinal contents were washed, sieved and examined microscopically for the presence of helminths, which were identified and counted. Results: Intestines from 108 wild dogs, 7 foxes and 18 Fraser Island dingoes were examined. Echinococcus granulosus was only present in the wild dogs from Maroochy Shire (46.3%) with worm burdens of between 30 and 104,000. Other helminths included Spirometra erinacei, Dipylidium caninum, Taenia spp., Ancylostoma caninum and Toxocara canis. Two specimens of a trematode (Haplorchinae sp.) usually found infecting fish and seabirds were recovered from a Fraser Island dingo. Conclusion: Dingoes on Fraser Island are not infected with E. granulosus and do not pose a hydatid disease public health risk to residents or visitors. However, wild dogs examined from the Maroochy Shire do present a potential hydatid disease public health risk.

Needs for applied climate education in agriculture

This paper reports on a purposive survey study which aimed to identify needs for the development, delivery and evaluation of applied climate education for targeted groups, to improve knowledge and skills to better manage under variable climatic conditions. The survey sample consisted of 80 producers and other industry stakeholders in Australia (including representatives from consulting, agricultural extension and agricultural education sectors), with a 58% response rate to the survey. The survey included an assessment of (i) knowledge levels of the Southern Oscillation Index and sea surface temperatures, and (ii) skill and ability in interpreting weather and climate parameters. Results showed that despite many of the respondents having more than 20 years experience in their industry, the only formal climate education or training undertaken by most was a 1-day workshop. Over 80% of the applied climate skills listed in the survey were regarded by respondents as essential or important, but only 42% of educators, 30% of consultants and 28% of producers rated themselves as competent in applying such skills. Essential skills were deemed as those that would enable respondents or their clients to be better prepared for the next extended wet or dry meteorological event, and improved capability in identifying and capitalising on key decision points from climate information and a seasonal climate outlook. The complex issue of forecast accuracy is a confounding obstacle for many in the application of climate information and forecasts in management. Addressing this problem by describing forecast 'limitations and skill' can help to overcome this problem. The survey also highlighted specific climatic tactical and strategic information collated from grazing, cropping and agribusiness enterprises, and showed the value of such information from a users perspective.

Is night-time wind direction important to best practice wild dog trapping and baiting?

We discovered a significant bias for wild dog scent station spoor (scats and scratches) to be positioned on the north-easterly side of roads and intersections. Counts of this spoor, 50 metres in each direction of north-south and east-west intersections were made in state forests near Roma in southwest Queensland, Cecil Plains on the Darling Downs and Maryborough on the coast during mating season in April/May 2007. While 51% of 190 and 83% of 120 scent station spoor were located on the north-eastern sector of the intersections at Cecil Plains and Roma respectively, spoor were more evenly distributed across all four sectors at Maryborough (n=47).

Are we focussing wild dog control the wrong time of the year and going about it the wrong way?

Our evaluation of the predation of calves by wild dogs in the 1990s found that the number of calves killed and frequency of years that calf losses occurred, is higher in baited areas compared to adjoining, non-baited areas of similar size. Calf losses were highest with poor seasonal conditions, low prey numbers and where baited areas were re-colonised by wild dogs soon after baiting. We monitored wild dog “activity” before and after 35 baiting programs in southwest, central west and far north Queensland between 1994 and 2006 and found change in activity depends on the timing of the baiting. Baiting programs conducted between October and April show an increase in dog activity post-baiting (average increase of 219.1%, SEM 100.9, n=9, for programs conducted in October and November; an increase of 82.5%, SEM 54.5, n=7 for programs conducted in March and April; and a decrease in activity of 46.5%, SEM 10.2, n=19 for programs conducted between May and September). We monitored the seasonal activity and dispersal of wild dogs fitted with satellite transmitters 2006 to present. We have found that: • Activity of breeding males and females, whilst rearing and nurturing pups, is focussed around the den between July to September and away from areas of human activity. Activity of breeding groups appears to avoid locations of human activity until juveniles become independent (around late November). • While independent and solitary yearlings often have unstable, elliptically-shaped territories in less favourable areas, members of breeding groups have territories that appear seasonally stable and circular located in more favourable habitats. • Extra-territorial forays of solitary yearlings can be huge, in excess of 200 km. The largest forays we have monitored have occurred when the activity of pack members is focussed around rearing pups and juveniles (August to November). • Where wild dogs have dispersed or had significant territorial expansion, it has occurred within days of baiting programs and onto recently baited properties. • The wild dogs we have tracked have followed netting barrier fences for hundreds of kilometres and lived adjacent to or bypassed numerous grids in the barrier. Based on these studies, we conclude that a proportion of the perceived decline in dog activity between May and September, post baiting, is due to a decline in dog activity in areas associated with human activity. The increase in dog activity post-baiting between October and May (and increased calf predation on baited properties) is likely caused by wild dogs dispersing (juveniles and yearlings) or expanding (adults) their territory into baited, now ‘vacant’, areas. We hypothesise that baiting programs should be focussed in summer and autumn commencing late November as soon as juveniles become independent of adults. We also hypothesise that instead of large, annual or semi-annual baiting programs, laying the same number of baits over 4-6 weeks may be more effective. These hypotheses need to be tested through an adaptive management project.

Post capture necrosis of the feet caused by soft-catch traps – incidents, cause and prevention.

When recapturing satellite collared wild dogs that had been trapped one month previous in padded foothold traps, we noticed varying degrees of pitting on the pads of their trapped paw. Veterinary advice, based on images taken of the injuries, suggests that the necrosis was caused by vascular compromise. Five of six dingoes we recaptured had varying degrees of necrosis restricted only to the trapped foot and ranging from single 5 mm holes to 25% sections of the toe pads missing or deformed, including loss of nails. The traps used were rubber-padded, two–coiled, Victor Soft Catch #3 traps. The springs are not standard Victor springs but were Beefer springs; these modifications slightly increase trap speed and the jaw pressure on the trapped foot. Despite this modification the spring pressure is still relatively mild in comparison to conventional long spring or four-coiled wild dog traps. The five wild dogs developing necrosis were trapped in November 2006 at 5-6 months of age. Traps were checked each morning so the dogs were unlikely to have been restrained in the trap for more than 12 hours. All dogs exhibited a small degree of paw damage at capture which presented itself as a swollen paw and compression at the capture point. In contrast, eight wild dogs, 7-8 month-old, were captured two months later in February. Upon their release, on advice from a veterinarian, we massaged the trapped foot to get blood flow back in to the foot and applied a bruise treatment (Heparinoid 8.33 mg/ml) to assist restoring blood flow. These animals were subsequently recaptured several months later and showed no signs of necrosis. While post-capture foot injuries are unlikely to be an issue in conventional control programs where the animal is immediately destroyed, caution needs to be used when releasing accidentally captured domestic dogs or research animals captured in rubber-padded traps. We have demonstrated that 7-8 month old dogs can be trapped and released without any evidence of subsequent necrosis following minimal veterinary treatment. We suspect that the rubber padding on traps may increase the tourniquet effect by wrapping around the paw and recommend the evaluation of offset laminated steel jaw traps as an alternative. Offset laminated steel jaw traps have been shown to be relatively humane producing as few foot injuries as rubber-jawed traps.

Tadpole mouthpart depigmentation as an accurate indicator of chytridiomycosis, an emerging disease of amphibians

Chytridiomycosis is an emerging infectious disease of amphibians caused by the fungal pathogen Batrachochytrium dendrobatidis, and its role in causing population declines and species extinctions worldwide has created an urgent need for methods to detect it. Several reports indicate that in anurans chytridiomycosis can cause the depigmentation of tadpole tnouthparts, but the accuracy of using depigmentation to determine disease status remains uncertain. Our objective was to determine for the Mountain Yellow-legged Frog (Rana muscosa) whether visual inspections of the extent of tadpole mouthpart depigmentation could be used to accurately categorize individual tadpoles or R. muscosa populations as B. dendrobatidis-positive or negative. This was accomplished by assessing the degree of mouthpart depigmentation in tadpoles of known disease status (based on PCR assays). The depigmentation of R. muscosa tadpole mouthparts was associated with the presence of B. dendrobatidis, and this association was particularly strong for upper jaw sheaths. Using a rule that classifies tadpoles with upper jaw sheaths that are 100% pigmented as uninfected and those with jaw sheaths that are <100% pigmented as infected resulted in the infection status of 86% of the tadpoles being correctly classified. By applying this rule to jaw sheath pigmentation scores averaged across all tadpoles inspected per site, we were able to correctly categorize the infection status of 92% of the study populations. Similar research on additional anurans is critically needed to determine how broadly applicable our results for R. muscosa are to other species.

Scat contents of the spotted-tailed quoll Dasyurus maculatus in the New England gorges, north-eastern New South Wales

In 1313 scats of the spotted-tailed quoll Dasyurus maculatus, collected over 5 years from the gorge country of north-eastern New South Wales, the most frequent and abundant items were derived from mammals and a restricted set of insect orders. These quolls also ate river-associated items: waterbirds, eels, crayfish, aquatic molluscs and even frogs. Macropods contributed most of the mammal items, with possums, gliders and rodents also being common. Some food, particularly from macropods and lagomorphs, had been scavenged (as shown by fly larvae). The most frequent invertebrates were three orders of generally large insects Coleoptera, Hemiptera and Orthoptera, which were most frequent in summer and almost absent in winter scats. Monthly mean numbers of rodent and small dasyurid items per scat were inversely related to these large insects in scats. The numbers of reptile items were inversely related to the numbers of mammal (especially arboreal and small terrestrial mammal) items per scat, thus types of items interacted in their occurrences in monthly scat samples. Frequencies of most vertebrate items showed no seasonal, but much year-to-year, variation. This quoll population ate four main types of items, each requiring different skills to obtain: they hunted arboreal marsupials (possibly up trees), terrestrial small mammals and reptiles (on the ground), and seasonally available large insects (on trees or the ground), and scavenged carcases, mostly of large mammals but also birds and fishes (wherever they could find them).

Assessing the taxonomic status of dingoes Canis familiaris for conservation

1. The conservation status of the dingo Canis familiaris dingo is threatened by hybridization with the domestic dog C. familiaris familiaris. A practical method that can estimate the different levels of hybridization in the field is urgently required so that animals below a specific threshold of dingo ancestry (e.g. 1/4 or 1/2 dingoes) can reliably be identified and removed from dingo populations. 2. Skull morphology has been traditionally used to assess dingo purity, but this method does not discriminate between the different levels of dingo ancestry in hybrids. Furthermore, measurements can only be reliably taken from the skulls of dead animals. 3. Methods based on the analysis of variation in DNA are able to discriminate between the different levels of hybridization, but the validity of this method has been questioned because the materials currently used as a reference for dingoes are from captive animals of unproven genetic purity. The use of pre-European materials would improve the accuracy of this method, but suitable material has not been found in sufficient quantity to develop a reliable reference population. Furthermore, current methods based on DNA are impractical for the field-based discrimination of hybrids because samples require laboratory analysis. 4. Coat colour has also been used to estimate the extent of hybridization and is possibly the most practical method to apply in the field. However, this method may not be as powerful as genetic or morphological analyses because some hybrids (e.g. Australian cattle dog × dingo) are similar to dingoes in coat colour and body form. This problem may be alleviated by using additional visual characteristics such as the presence/absence of ticking and white markings.

A strategic approach to mitigating the impacts of wild canids: proposed activities of the Invasive Animals Cooperative Research Centre

Wild canids (wild dogs and European red foxes) cause substantial losses to Australian livestock industries and environmental values. Both species are actively managed as pests to livestock production. Contemporaneously, the dingo proportion of the wild dog population, being considered native, is protected in areas designated for wildlife conservation. Wild dogs particularly affect sheep and goat production because of the behavioural responses of domestic sheep and goats to attack, and the flexible hunting tactics of wild dogs. Predation of calves, although less common, is now more economically important because of recent changes in commodity prices. Although sometimes affecting lambing and kidding rates, foxes cause fewer problems to livestock producers but have substantial impacts on environmental values, affecting the survival of small to medium-sized native fauna and affecting plant biodiversity by spreading weeds. Canid management in Australia relies heavily on the use of compound 1080-poisoned baits that can be applied aerially or by ground. Exclusion fencing, trapping, shooting, livestock-guarding animals and predator calling with shooting are also used. The new Invasive Animals Cooperative Research Centre has 40 partners representing private and public land managers, universities, and training, research and development organisations. One of the major objectives of the new IACRC is to apply a strategic approach in order to reduce the impacts of wild canids on agricultural and environmental values in Australia by 10%. In this paper, the impacts, ecology and management of wild canids in Australia are briefly reviewed and the first cooperative projects that will address IACRC objectives for improving wild dog management are outlined.

Resistance to root-lesion nematodes (Pratylenchus thornei and P-neglectus) in synthetic hexaploid wheats and their durum and Aegilops tauschii parents.

Root-lesion nematodes (Pratylenchus thornei Sher and Allen and P. neglectus (Rensch) Filipijev and Schuurmans Stekhoven) cause substantial yield loss to wheat crops in the northern grain region of Australia. Resistance to P. thornei for use in wheat breeding programs was sought among synthetic hexaploid wheats (2n= 6x = 42, AABBDD) produced through hybridisations of Triticum turgidum L. subsp. durum (Desf.) Husn (2n= 4x = 28, AABB) with Aegilops tauschii Coss. (2n= 2x = 14, DD). Resistance was determined for the synthetic hexaploid wheats and their durum and Ae. tauschii parents from the numbers of nematodes in the roots of plants grown for 16 weeks in pots of pasteurised soil inoculated with P. thornei. Fifty-nine (32%) of 186 accessions of synthetic hexaploid wheats had lower numbers of nematodes than Gatcher Selection 50a (GS50a), a partially resistant bread wheat. Greater frequencies of partial resistance were present in the durum parents (72% of 39 lines having lower nematode numbers than GS50a) and in the Ae. tauschii parents (55% of 53 lines). The 59 synthetic hexaploids were re-tested in a second experiment along with their parents. In a third experiment, 11 resistant synthetic hexaploid wheats and their F-1 hybrids with Janz, a susceptible bread wheat, were tested and the F(1)s were found to give nematode counts intermediate between the respective two parents. Synthetic hexaploid wheats with higher levels of resistance resulted from hybridisations where both the durum and Ae. tauschii parents were partially resistant, rather than where only one parent was partially resistant. These results suggest that resistance to P. thornei in synthetic hexaploid wheats is polygenic, with resistances located both in the D genome from Ae. tauschii and in the A and/or B genomes from durum. Five synthetic hexaploid wheats were selected for further study on the basis of (1) a high level of resistance to P. thornei of the synthetic hexaploid wheats and of both their durum and Ae. tauschii parents, (2) being representative of both Australian and CIMMYT (International Maize and Wheat Improvement Centre) durums, and (3) being representative of the morphological subspecies and varieties of Ae. tauschii. These 5 synthetic hexaploid wheats were also shown to be resistant to P. neglectus, whereas GS50a and 2 P. thornei-resistant derivatives were quite susceptible. Results of P. thornei resistance of F(1)s and F(2)s from a half diallel of these 5 synthetic hexaploid wheats, GS50a, and Janz from another study indicate polygenic additive resistance and better general combining ability for the synthetic hexaploid wheats than for GS50a. Published molecular marker studies on a doubled haploid population between the synthetic hexaploid wheat with best general combining ability (CPI133872) and Janz have shown quantitative trait loci for resistance located in all 3 genomes. Synthetic hexaploid wheats offer a convenient way of introgressing new resistances to P. thornei and P. neglectus from both durum and Ae. tauschii into commercial bread wheats.

Clinical signs and duration of cyanide toxicosis delivered by the M-44 ejector in wild dogs

Sodium cyanide poison is potentially a more humane method to control wild dogs than sodium fluoroacetate (1080) poison. This study quantified the clinical signs and duration of cyanide toxicosis delivered by the M-44 ejector. The device delivered a nominal 0.88 g of sodium cyanide, which caused the animal to loose the menace reflex in a mean of 43 s, and the animal was assumed to have undergone cerebral hypoxia after the last visible breath. The mean time to cerebral hypoxia was 156 s for a vertical pull and 434 s for a side pull. The difference was possibly because some cyanide may be lost in a side pull. There were three distinct phases of cyanide toxicosis: the initial phase was characterised by head shaking, panting and salivation; the immobilisation phase by incontinence, ataxia and loss of the righting reflex; and the cerebral hypoxia phase by a tetanic seizure. Clinical signs that were exhibited in more than one phase of cyanide toxicosis included retching, agonal breathing, vocalisation, vomiting, altered levels of ocular reflex, leg paddling, tonic muscular spasms, respiratory distress and muscle fasciculations of the muzzle.

Wild dogma: An examination of recent "evidence" for dingo regulation of invasive mesopredator release in Australia

There is growing interest in the role that apex predators play in shaping terrestrial ecosystems and maintaining trophic cascades. In line with the mesopredator release hypothesis, Australian dingoes (Canis lupus dingo and hybrids) are assumed by many to regulate the abundance of invasive mesopredators, such as red foxes Vulpes vulpes and feral cats Felis catus, thereby providing indirect benefits to various threatened vertebrates. Several recent papers have claimed to provide evidence for the biodiversity benefits of dingoes in this way. Nevertheless, in this paper we highlight several critical weaknesses in the methodological approaches used in many of these reports, including lack of consideration for seasonal and habitat differences in activity, the complication of simple track-based indices by incorporating difficult-to-meet assumptions, and a reduction in sensitivity for assessing populations by using binary measures rather than potentially continuous measures. Of the 20 studies reviewed, 15 of them (75%) contained serious methodological flaws, which may partly explain the inconclusive nature of the literature nvestigating interactions between invasive Australian predators. We therefore assert that most of the “growing body of evidence” for mesopredator release is merely an inconclusive growing body of literature only. We encourage those interested in studying the ecological roles of dingoes relative to invasive mesopredators and native prey species to account for the factors we identify, and caution the value of studies that have not done so.