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.

Welfare outcomes for 3- and 6-month-old beef calves in a tropical environment castrated surgically or by applying rubber rings

Castration of cattle using rubber rings is becoming increasingly popular due to the perceived ease of the procedure and greater operator safety when compared with surgical castration. Few comparative studies have investigated the effects of different castration methods and calf age on welfare outcomes, particularly in a tropical environment. Thirty Belmont Red (a tropically adapted breed), 3-month-old (liveweight 71–119 kg) and 30, 6-month-old (liveweight 141–189 kg) calves were assigned to a two age × three castration (surgical, ring and sham) treatment factorial study (Surg3, Surg6, Ring3, Ring6, Sham3 and Sham6, n = 10 for each treatment group). Welfare outcomes were assessed post-castration using: behaviour for 2 weeks; blood parameters (cortisol and haptoglobin concentrations) to 4 weeks; wound healing to 5 weeks; and liveweights to 6 weeks. More Surg calves struggled during castration compared with Sham and Ring (P < 0.05, 90 ± 7% vs. 20 ± 9% and 24 ± 10%) and performed more struggles (1.9 ± 0.2, 1.1 ± 0.3 and 1.1 ± 0.3 for Surg, Sham and Ring, respectively), suggesting that surgical castration caused most pain during performance of the procedure. A significant (P < 0.05) time × castration method × age interaction for plasma cortisol revealed that concentrations decreased most rapidly in Sham; the Ring6 calves failed to show reduced cortisol concentrations at 2 h post-castration, unlike other treatment groups. By 7 h post-castration, all treatment groups had similar concentrations. A significant (P < 0.01) interaction between time and castration method showed that haptoglobin concentrations increased slightly to 0.89 and 0.84 mg/mL for Surg and Ring, respectively over the first 3 days post-castration. Concentrations for Surg then decreased to levels similar to Sham by day 21 and, although concentrations for Ring decreased on day 7 to 0.76 mg/mL, they increased significantly on day 14 to 0.97 mg/mL before reducing to concentrations similar to the other groups (0.66 mg/mL) by day 21. Significantly (P < 0.05) more of the wounds of the 3-month compared with the 6-month calves scored as ‘healed’ at day 7 (74% vs. 39%), while more (P = 0.062) of the Surg than Ring scored as ‘healed’ at day 21 (60% vs. 29%). At day 14 there were significantly (P < 0.05) fewer healed wounds in Ring6 compared with other treatment groups (13% vs. 40–60%). Liveweight gain was significantly (P < 0.05) greater in 3-month (0.53 kg/day) than in 6-month calves (0.44 kg/day) and in Sham calves (P < 0.001, 0.54 kg/day), than in Ring (0.44 kg/day) and Surg (0.48 kg/day) calves. Overall, welfare outcomes were slightly better for Surg than Ring calves due to reduced inflammation and faster wound healing, with little difference between age groups.

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.

Welfare outcomes for 3- and 6-month-old beef calves in a tropical environment castrated surgically or by applying rubber rings

Castration of cattle using rubber rings is becoming increasingly popular due to the perceived ease of the procedure and greater operator safety when compared with surgical castration. Few comparative studies have investigated the effects of different castration methods and calf age on welfare outcomes, particularly in a tropical environment. Thirty Belmont Red (a tropically adapted breed), 3-month-old (liveweight 71–119 kg) and 30, 6-month-old (liveweight 141–189 kg) calves were assigned to a two age × three castration (surgical, ring and sham) treatment factorial study (Surg3, Surg6, Ring3, Ring6, Sham3 and Sham6, n = 10 for each treatment group). Welfare outcomes were assessed post-castration using: behaviour for 2 weeks; blood parameters (cortisol and haptoglobin concentrations) to 4 weeks; wound healing to 5 weeks; and liveweights to 6 weeks. More Surg calves struggled during castration compared with Sham and Ring (P < 0.05, 90 ± 7% vs. 20 ± 9% and 24 ± 10%) and performed more struggles (1.9 ± 0.2, 1.1 ± 0.3 and 1.1 ± 0.3 for Surg, Sham and Ring, respectively), suggesting that surgical castration caused most pain during performance of the procedure. A significant (P < 0.05) time × castration method × age interaction for plasma cortisol revealed that concentrations decreased most rapidly in Sham; the Ring6 calves failed to show reduced cortisol concentrations at 2 h post-castration, unlike other treatment groups. By 7 h post-castration, all treatment groups had similar concentrations. A significant (P < 0.01) interaction between time and castration method showed that haptoglobin concentrations increased slightly to 0.89 and 0.84 mg/mL for Surg and Ring, respectively over the first 3 days post-castration. Concentrations for Surg then decreased to levels similar to Sham by day 21 and, although concentrations for Ring decreased on day 7 to 0.76 mg/mL, they increased significantly on day 14 to 0.97 mg/mL before reducing to concentrations similar to the other groups (0.66 mg/mL) by day 21. Significantly (P < 0.05) more of the wounds of the 3-month compared with the 6-month calves scored as ‘healed’ at day 7 (74% vs. 39%), while more (P = 0.062) of the Surg than Ring scored as ‘healed’ at day 21 (60% vs. 29%). At day 14 there were significantly (P < 0.05) fewer healed wounds in Ring6 compared with other treatment groups (13% vs. 40–60%). Liveweight gain was significantly (P < 0.05) greater in 3-month (0.53 kg/day) than in 6-month calves (0.44 kg/day) and in Sham calves (P < 0.001, 0.54 kg/day), than in Ring (0.44 kg/day) and Surg (0.48 kg/day) calves. Overall, welfare outcomes were slightly better for Surg than Ring calves due to reduced inflammation and faster wound healing, with little difference between age groups.

Integrating Rapid Phenotyping and Speed Breeding to Improve Stay-Green and Root Adaptation of Wheat in Changing, Water-Limited, Australian Environments

Temperatures have increased and in-crop rainfall decreased over recent decades in many parts of the Australian wheat cropping region. With these trends set to continue or intensify, improving crop adaptation in the face of climate change is particularly urgent in this, already drought-prone, cropping region. Importantly, improved performance under water-limitation must be achieved while retaining yield potential during more favourable seasons. A multi-trait-based approach to improve wheat yield and yield stability in the face of water-limitation and heat has been instigated in northern Australia using novel phenotyping techniques and a nested association mapping (NAM) approach. An innovative laboratory technique allows rapid root trait screening of hundreds of lines. Using soil grown seedlings, the method offers significant advantages over many other lab-based techniques. Another recently developed method allows novel stay-green traits to be quantified objectively for hundreds of genotypes in standard field trial plots. Field trials in multiple locations and seasons allow evaluation of targeted trait values and identification of superior germplasm. Traits, including yield and yield components are measured for hundreds of NAM lines in rain fed environments under various levels of water-limitation. To rapidly generate lines of interest, the University of Queensland “speed breeding” method is being employed, allowing up to 7 plant generations per annum. A NAM population of over 1000 wheat recombinant inbred lines has been progressed to the F5 generation within 18 months. Genotyping the NAM lines with the genome-wide DArTseq molecular marker system provides up to 40,000 markers. They are now being used for association mapping to validate QTL previously identified in bi-parental populations and to identify novel QTL for stay-green and root traits. We believe that combining the latest techniques in physiology, phenotyping, genetics and breeding will increase genetic progress toward improved adaptation to water-limited environments.

Sirex woodwasp range expansion in Australia: performance and parasitism on two commercial pine species

Sirex woodwasp was detected in Queensland in 2009 and rapidly established in softwood plantations (Pinus radiata and P. taeda) in southern border regions. Biocontrol inoculations of Deladenus siricidicola began soon after, and adults were monitored to assess the success of the programme. Wasp size, sex ratios, emergence phenology and nematode parasitism rates were recorded, along with the assessment of wild-caught females. Patterns varied within and among seasons, but overall, P. taeda appeared to be a less suitable host than P. radiata, producing smaller adults, lower fat body content and fewer females. Sirex emerging from P. taeda also showed lower levels of nematode parasitism, possibly due to interactions with the more abundant blue-stain fungus in this host. Sirex adults generally emerged between November and March, with distinct peaks in January and March, separated by a marked drop in emergence in early February. Temperature provided the best correlate of seasonal emergence, with fortnights with higher mean minimum temperatures having higher numbers of Sirex emerging. This has implications for the anticipated northward spread of Sirex into sub-tropical coastal plantation regions. Following four seasons of inundative release of nematodes in Queensland, parasitism rates remain low and have resulted in only partial sterilization of infected females.

Polybridge Technical Report

This study examined the physical and chemical properties of a novel, fully-recirculated prawn and polychaete production system that incorporated polychaete-assisted sand filters (PASF). The aims were to assess and demonstrate the potential of this system for industrialisation, and to provide optimisations for wastewater treatment by PASF. Two successive seasons were studied at commercially-relevant scales in a prototype system constructed at the Bribie Island Research Centre in Southeast Queensland. The project produced over 5.4 tonnes of high quality black tiger prawns at rates up to 9.9 tonnes per hectare, with feed conversion of up to 1.1. Additionally, the project produced about 930 kg of high value polychaete biomass at rates up to 1.5 kg per square metre of PASF, with the worms feeding predominantly on waste nutrients. Importantly, this closed production system demonstrated rapid growth of healthy prawns at commercially relevant production levels, using methods that appear feasible for application at large scale. Deeper (23 cm) PASF beds provided similar but more reliable wastewater treatment efficacies compared with shallower (13 cm) beds, but did not demonstrate significantly greater polychaete productivity than (easier to harvest) shallow beds. The nutrient dynamics associated with seasonal and tidal operations of the system were studied in detail, providing technical and practical insights into how PASF could be optimised for the mitigation of nutrient discharge. The study also highlighted some of the other important advantages of this integrated system, including low sludge production, no water discharge during the culture phase, high ecosystem health, good prospects for biosecurity controls, and the sustainable production of a fishery-limited resource (polychaetes) that may be essential for the expansion of prawn farming industries throughout the world. Regarding nutrient discharge from this prototype mariculture system, when PASF was operating correctly it proved feasible to have no water (or nutrient) discharge during the entire prawn growing season. However, the final drain harvest and emptying of ponds that is necessary at the end of the prawn farming season released 58.4 kg ha-1 of nitrogen and 6 kg ha-1 of phosphorus (in Season 2). Whilst this is well below (i.e., one-third to one-half of) the current load-based licencing conditions for many prawn farms in Australia, the levels of nitrogen and chlorophyll a in the ponds remained higher than the more-stringent maximum limits at the Bribie Island study site. Zero-net-nutrient discharge was not achieved, but waste nutrients were low where 5.91 kg of nitrogen and 0.61 kg of phosphorus was discharged per tonne of prawns produced. This was from a system that deployed PASF at 14.4% of total ponded farm area which treated an average of 5.8% of pond water daily and did not use settlement ponds or other natural or artificial water remediation systems. Four supplemental appendices complement this research by studying several additional aspects that are central to the industrialisation of PASF. The first details an economic model and decision tool which allows potential users to interactively assess construction and operational variables of PASF at different scales. The second provides the qualitative results of a prawn maturation trial conducted collaboratively with the Commonwealth Scientific and Industrial Research Organisation (CSIRO) to assess dietary inclusions of PASF-produced worms. The third provides the reproductive results from industry-based assessments of prawn broodstock produced using PASF. And the fourth appendix provides detailed elemental and nutritional analyses of bacterial biofilm produced by PASF and assesses its potential to improve the growth of prawns in recirculated culture systems.