Lecture Flashcards
What is the most widespread and destructive defoliator of coniferous forests in western NA?
- Western Spruce Budworm
What does WSB feed on?
- Larvae emerge after overwintering at same time as new spruce buds emerge and feed on them
- Coevolution to emerge at same time as buds develop on trees
What is the tree scale habitat of WSB?
- Bud flush: timing and abundance
- Size of Crown and available foliage (puffy = lots of food)
What is the stand scale factors of WSB?
- Species composition (homogeneous good for connectivity, hetero bad)
- Stand structure (Vertical heterogeneity and multiple age classes good, spin down from top and reach more food sources of lower tree tops)
- Site quality
- Standscale alone cannot fully explain spatiotemporal patterns of insect outbreaks
What could control the expansion of distinct and randomly distributed infestation patches to more continuous landscape level outbreaks?
- Landscape-level factors:
- Composition, configuration of host populations (connectivity, abundance)
- Adult moth dispersal and pred-prey interactions (dense forest decreases bird predators and increases dispersal)
What could control the expansion of distinct and randomly distributed infestation patches to more continuous landscape level outbreaks?
- Regional Factors:
- Physiography
- Climate (moisture deficits, warm dry conditions to disperse and survive, autumn precipitation)
- Drought year before outbreak causes outbreak initiation due to survivability and dispersal as well as increased nutrition in hosts from stress forcing more sugar into needles
What are the most important infestation predictors?
- Proximity to infestations in the previous year
- Landscape-scale host abundance
- Dry autumn conditions
What scale is the most important to manage for WSB outbreaks?
- Landscape because these are the factors that contribute most to the outbreaks
Pattern and fire
Species compostition and structure, including fuel amounts, size classes and arrangement
- Affects how a fire will burn
Low severity fire and where in BC does it occur
- Ecological effects minimal, species are adapted
- Fire consumes surface fuel, not into large tree canopy
- Dry forest that experiences frequent fire (except for the modern suppression)
- Interior BC, Douglas Fir, Ponderosa Pine
Severity
- Ecological effects above and below ground of a disturbance
High severity fire and where in BC does this occur
- Ecological effects profound
- Big highly flammable trees not adapted to fire, tree crowns burn
- Coastal and northern forests
Mixed severity fire and where does it occur
- Some patches of high and of low severity
- Dry and northern forests
- More common/dominant in NA than previously thought
Patch size of low severity fires
- Openings created by fire within which post-fire regeneration occurs
- Low-severity fire regime generally has small patches in larger matrix of homogeneous forest cover
Ponderosa pine forests and fire
- Open forest so fire moves quickly across landscape, consumes fuel on surface, doesn’t burn in one place for long, bark resistant to fire, not very flammable
- Small patches with homogeneous matrix
Patch size of mixed severity fires
- Results in both large and small patches from areas of low, moderate, and high severity fire
Patch size of high severity fires
- Small patches of isolated extreme fire activity or very large patches that an remain treeless for up to a century after the fire
High elevations and fire
- not well adapted to fires
- Regeneration time may be long
Landscape Metrics: Patch Edge
- Calculate edge index by measuring whole perimeter of disturbance (fire), including the undisturbed islands within patch
- Compute ratio between total perimeter and the perimeter of a circle of same size to get edge index
- Can be adapted to measure/compare edges around areas affected by different severities (fire)
High vs. low edge index
- High: very impactful for regeneration
- Low: less edge and less recolonization options and longer regeneration time
Which is easier to define, edges of high or low severity fire patches?
- High severity is more prominent
- Low severity is diffuse and difficult to define
What shape does a patch of wind driven fire take?
- Oblong or oval
Which fire regime has higher edge index value?
- Mixed
- Indicates greater landscape level heterogeneity in pattern
Describe fire characteristics of high vs. low severity or index
- Low severity: small patches, low edge index, highly similar pre-post fire
- High severity: large variable patches, depends on topography, moderate edge, high index, low pre-post fire similarity
- Mixed severity: moderate patch size, high edge, moderate pre-post fire similarity
What factors could contribute to fires with greater edge indices?
- Weather: Wind driven fires, larger fires w/ longer burn length and greater topographic variability
What are the effects of patch edge with fire?
- Microclimates
- Birds like edge environments (increase MBP and WSB predators)
- Quantifiably measure landscape and apply to other features
- Reduce connectivity and host abundance for pests?
Measures of landscape connectivity
- Spatial graphs
- Network analysis
- Social network models
Metrics of landscape composition
- Refers to cover types in an area and how much of each class is present
- Not spatially explicit
- Fraction occupied
- Diversity and dominance
Measures of spatial configuration
- Quantitative description of the spatial arrangement of cover types on the landscape
- Edge length and edge density
- Contagion
- Patch-based metrics
Fraction occupied
- Calc proportion of landscape that is occupied by each cover type
- Estimate proportion by counting number of grid cells over landscape that are occupied by a cover type, then divide by total number of grid cells
Diversity and Dominance
- Based on relative abundance of each cover type
- Inversely related, usually only 1 reported because implies the other
- Diversity refers to how evenly the proportion of cover types are distributed
- Require at least 2 cover types
Normalized landscape diversity metric
- H = diversity
- pi = Proportion of landscape occupied by cover type i
- s = number of cover types present
- H = (negative sum of pi ln(pi))/ ln (s)
What is a limitation of H (diversity) and D (dominance) metric
- Metrics tied to proportions, not qualitative aspects of landscape
- Same H and D may occur for landscapes that have a different proportion of cover (i.e. 10% forest and 90% agriculture = same H and D as 10% agriculture and 90% forest)
Study or concept when diversity or dominance metric would be useful to make comparisons?
- MaMu and logged landscapes and old growth
- Sea otters and kelp forests
- Urban environments of apartments vs. retail vs. park etc. (compare w/ pollution and calc with index of dominance leads to most pollution
- Compare species invasion w/ native diversity and which areas more susceptible to invasion
What is often excluded in the calculation of edges?
- The perimeter of the map
- Don’t know what is happening outside of map, can’t use
Edge density
- Calculated by summing edge length and dividing by map area
- Can be tallied total or by cover type
Contagion Metric
- Uses adjacency info and distinguishes btwn overall landscape patterns that are clumped or dissected
- Probabilities of adjacency, how likely is it to occur
- Sensitive to fine-scale spatial distribution of cover types
- Adjacency of similar and different cover types
Contagion Metric pre-calculation
- qi,j = ni,j/ni
- qi,j = probability of adjacency
- ni = number of grid cells of cover type i
- ni,j = number of instances when cover type i is adjacent to cover type j
Contagion Metric calculation
C = 1 + sum of i * sum of j [(piqij)ln(pi*qij)]/2ln(s)
- s = area of inquiry
Patch-based metrics
- Patch number, size, perimeter and shape
- Report for individual cover types
- Frequency distributions of numbers of patches and the mean, median and SD of patch size
Perimeter to area ratios
- Index of shape complexity
- High P/A = complex shape
- Low P/A = compact and simple shape
- Sensitive to patch size
Largest patch index
- Relate to fragmentation of a given cover type
- Calculates the size of the largest patch relative to the maximum size possible if the cover type occurred as a single patch
- Straightforward way to characterize landscape and compare btwn study areas
Largest patch index calc
- When index = 1?
LPIi = LCi/(pixmxn)
- LCi = size of largest patch of habitat i
- pi = proportion of landscape occupied by habitat i
- m x n = gives size of landscape of m row and n column
- if all cover i occurs as single patch, value of index = 1 (complete connectivity)
- When dispersed, index approaches 0
Patch Isolation
- Degree to which patches are isolated from other same cover patches
- Connectivity and fragmentation (habitat-use pattern studies)
- Mean inter patch distance
- Could be applied for one cover type or for more (animal that requires 2 habitats in close proximity
Mean inter patch distance
- Distance from centre of one patch to the centre of the next nearest patch
Proximity index
- Relative isolation of patches
- Low = isolation, high = well connected patches
Proximity index calculation
PXi = sum of (Sk/nk)
- PXi = proximity index for local patch i
- Specific search area
- Sk = area of patch k w/in search area
- nk = nearest neighbour distance btwn the grid cell of the focal patch and the nearest grid cell of patch k
Issue of scale with patch-based metrics?
- perimeter and area will change by scale, zoom in and scale changes (fine-scale = more detail discernible and more perimeter/area/complexity than at broad scale)
- Only real fix is to report scale of research when writing about
Fractals
- Self-replicating data
- Small shapes seen at larger scale in replication
- Never ending
- Shape made of parts of the whole
- Mathematics of fractals thought to be able to fix scale issue of landscape ecology (shape of fine applies to large)
Importance of measures of landscape connectivity
- Some landscapes, cell-based approaches not good models, alternatives (dendritic, road density) could be used
- Fragmented habitats, degree that organisms can move among patches becomes more important
- Connectivity in conservation and reserve design has fostered a proliferation of connectivity metrics
Connectivity
- Structurally, based on habitat patterns and assumptions about organism dispersal, or functionally, based on where organisms actually move
- Property of entire landscape or of a particular habitat patch
- Degree to which landscape facilitates or impedes movement of organisms
- Difference between patch connectivity and landscape connectivity
Connectivity and MPB, what affects MPB functional connectivity?
- Important b/c poor dispersers
- Wind, pheromones to help guide beetles to good habitat, predators, birds, populations drive MPB movement through landscape
Example of climate change effect on connectivity
- Polar bears use ice to connect landscape
- More energy use to travel on landscape w/o ice
Graph theory
- Pattern across landscape as series of nodes (represent habitat patches) and links (connections among points/nodes)
- Used to examine connectivity, id pathways for dispersal/movement, and prioritize patches for conservation
Gamma index, spatial graphs
- gamma = L/Lmax = L/3(V-2)
- L = number of links
- V = number of nodes
- Values close to 0 = little connectivity, close to 1 = high connectivity
Which patch(es) should be prioritized?
- Focal b/c important for whole landscape, maybe prioritize other close patches but should also layer other habitat info such as water, predators, range etc.
- Must understand organisms and other landscape features (are close patches important or further important for range)
Network Analysis
- More complex than graph theory
- number of nodes, links, degree to which nodes are connected directly or indirectly
- Nodes can be weighted by area or habitat quality to better represent importance for connectivity
- What happens when a node or link is removed or added?
Social network models
- Incorporate movement into network analysis
- Theoretical dispersal vs. actually measuring movement
- Complex models
- Research frontier
Fragstats
- Software program designed to compute variety of landscape metrics for categorical map patterns
- Developed in 1995
- Patch-based, cell-based, surface metrics, structural and functional metrics, batch processing, sampling strategies,
Continuous vs. discrete data
- Categorical = discrete
- Landcape metrics effective with discrete datasets
- Spatial stats incorporates continuous data series
Tobler’s 1st law of Geography
- Everything is related to everything else, but near things are more related than distant things
2 approaches for spatial statistics
- Point pattern analysis, analyzes observed ‘events’ (nest locations, fire starts)
- Spatial autocorrelation and variography
Point Pattern Analysis
- Data is records of event-based spatial phenomena
- Ripley’s k-function
Ripley’s k-function
- K = 1/N2 * ASum of jSum of idIdij/wij
- N = number of points
- A is size of study area that contains points
- dij subset of distances less than id
- id range from min to max possible w/in A
- wij = edge correction required to avoid boundary effects
Interpreting ripleys-k
- Where the data departs from random theoretical line is where data is significant
- Above = clustered, below= dispersed
What are the landscape-level processes affecting conifer encroachment into grasslands?
- Climate, fire, grazing, biotic interactions
- Biotic interactions of facilitation among shade tolerant and intolerant tree species
Spatial autocorrelation and variography
- 2 widely used methods for characterizing spatial dependence or spatial structure in a variable as a function of its position in a landscape
Spatial autocorrelation
- Correlation of spatially, and temporally, distributed variables
- Near things are more related
- Association between features and determine distances at which relation is high and where it drops off
- Breaks assumption that things are random in statistics
Models
- Abstract representation of a system or a process
- Used in landscape ecology to represent at least one landscape pattern-process relationships of interest
- All models are wrong, but some are useful
Modelling tree-ring width to climate
- If significant strong correlation exists, build simple linear model where variability in tree ring is explained by climate
- Use model and time series of tree-ring to hindcast past climate
- Not perfect, but useful
Random maps
- Simple standard for landscape pattern
- Observed landscapes vs. replicate random maps reveal magnitude and significance of differences due to the structure of actual landscapes
Neutral landscape models, 2 categories of models
- Determine extent to which structural properties of landscapes (patch size, shape, edge connectivity autocorrelation) deviate from theoretical spatial distribution
- Predict how ecological processes such as animal movement, seed dispersal, gene flow, or fire spread are affected by landscape pattern
Landscape modelling: Conceptualization
- How does system work
- What are the entities that define the structure of the system
- What are key processes
Landscape modelling:
Formalization
- What are the state variables
- Mechanisms/constraints included/excluded
- Assumptions about system
- Spatial and temporal scale of model
Landscape modelling:
Implementation
- Form of model equations
- How will model be solved (language and platform)
Landscape modelling:
Parameterization
- Data needed to estimate all parameters and set the initial conditions of the model
Landscape modelling:
Verification
- Does model do what it was built to do?
- How accurate at representing true values?
Markov chains
- Change detection
- Observations of the state of a landscape at 2 time periods
- Transitional matrix
- Ex. Mamu and id habitat change over last 20 yrs
- Describe landcover change (remote sensing is useful)
What are the main things that models allow us to do?
- Handle complex multivariate relationships too complicated for human mind
- Project into future
- Id most influential driving factors, focus attention on the few things that matter most
- Id critical empirical information needs, focus on future research where it will do most good
- Id thresholds/system behaviour, inform management of where, when and how to avoid major impacts on system
- Explore ‘what if’ scenarios, cannot be done in real world
Natural range of variability
- Stems from ‘dynamic view’
- Ecological conditions, spatial and temporal variation in these conditions, that are relatively unaffected by people, w/in a period of time and geographical area appropriate to an expressed goal
4 steps for natural variability
- Obtain site based information
- Compile landcape history for sites
- Interpret landscape-scale disturbance regime from site specific data
- Develop landscape management plan using these findings
Management applications of natural variability
- Understanding and evaluating change
- Reference for setting general management goals
Target stand
- conditions from historical evidence
- Recreated historical stand structures
- selected an appropriate stand structure
- designed a silvicultural burn treatment
- implement treatment
Problems with natural range of variability approach
- Wrong scale for effective management
- Selected condition mostly arbitrary
- Difficult/inappropriate to treat entire landscape w/ one treatment
- Does not recognize inherent variability
- Past conditions may no longer exist (due to climate change for ex.)
- How do we know for sure what species assemblages existed before and have the species actually adapted to the new conditions
Examples of applications of natural range of variability
- Salmon forests and natural variability form tree rings of salmon and use of creek
- Caribou and use of snow fields to escape bugs by carbon dating feces trapped in ice to get inferences on population and snow use
What must be incorporated into any concept of landscape equilibrium?
- Disturbance needs to be incorporated in order for an valid definition of an equilibrium state
Landscape dynamics
- Considers spatial and temporal scales of disturbance and the resultant landscape dynamics
- Can be applied across range of scales
4 Factors of landscape dynamics
- Disturbance freq and return interval
- Rate of recovery from disturbance
- Size/spatial extent of landscape
- Reduce this to just temporal and spatial pattern
Temporal pattern (Landscape dynamics)
- Ratio of the disturbance interval to the recovery time (to mature stage)
- Disturbance interval longer than recovery time, system can recover before being disturbed again
- Disturbance interval and recovery time being equal (stable but variable)
- Disturbance interval shorter than recovery
Spatial pattern (landscape dynamics)
- Ratio of the size of disturbance relative to the size of landscape of interest
- Disturbances are large or small relative to landscape
What does the use of ratios in both temporal and spatial parameters permit in landscape dynamics?
- Permits comparisons of landscape across range of temporal and spatial scales
State-space diagram
- Temporal ratio vs. spatial ratio
- When close to 1 on spatial scale, the disturbance is more than landscape and therefore unstable
Landscape heterogeneity
- Animal behaviour
- Habitat selection
- Movement rates and trajectories
- Species interactions (pred-prey)
- Patterns and rate of spread of invasive species
Source
Habitat areas where local reproductive success is greater than mortality
Sink
- Poor habitats where local mortality exceeds reproductive success
Examples of important factors for wolf habitat?
- Vegetation type
- Deer populations
- Land ownership
- Road density
- Human population density
Study that analyzed wolf habitat and organisms and landscape pattern?
- Regional landscape analysis and prediction of favourable gray wolf habitat in northern great lakes region
- Mladenoff, Sickley et al.
Habitat selection
- Act of choosing combination of available abiotic and biotic elements that best fulfills the life-history needs of the organism
- Gather by trail cameras, hair snags, tracks (issues w/ when laid and how many individuals), radio collars
What is habitat?
- Continuos variable
- Can be mapped with lidar over landscape (and in 3D!)
Resource selection functions
- Describe what it is and how used in model of habitat for particular animal
Model that yields values proportional to the probability of use of a resource unit
- Often fit using linear models (continuous dataset)
- Elk in Yellowstone use more habitat in summer, less in winter where they cluster more (driven by predator prey interactions with wolves and resource availability)
- Elk moved frequently among preferred locations, not extend periods w/in preferred
- Elk move about landscape to avoid detection by wolves
How do organisms create landscape pattern? Example?
- Ecosystem engineers like beavers and hippopotamuses
- Hippos feed on aquatic veg, river and riparian habitat respond to the disturbance by creating more corridors and heterogeneity in landscape system as hippo use increases
Response of organism to landscape heterogeneity (5 points)
- Larger, more heterogeneous patches contain more species and more organisms
- Relative abundance of edge and interior habitat affects species diversity w/in a patch
- Characteristics of surrounding landscape can influence local populations in patch
- Effect of landscape composition on organisms is often stronger than effect of landscape configuration
- Corridor creation can both add habitat and promote movement (butterfly genetics, shape can have sig. effect on even fine-scale)
Edges and landscape heterogeneity
- Edges may be barriers (allow big mammals but not small) or filters to movement (filter male from female deer)
- Agents that alter mortality rates (less protection)
- Areas providing energetic subsidies or refuge (more insects for birds on edges)
- Regions where novel species interactions occur
Predator-Prey interactions
- How does heterogeneity mediate interactions between pred and prey
- Landscape pattern influences probability of prey encounters
- Each uses areas that give them advantages, reflective of each others usage
- Odds of elk encountering wolves was 1.3x greater in pine forest and 4.1x less in grasslands
- Cougars use large denser patches while deer use open areas away from edges
Stages of invasion
- Introduction (intentional or accidental)
- Colonization
- Establishment
- Dispersal
- Spatially distributed populations
- Invasive spread
Landscape effects of Introduction of invasive species
- Topographic effects on human land-use patterns may indirectly increase frequency of introductions
Landscape effects of Colonization of invasive species
- Spatial distribution of disturbed areas, safe sites, or resources required for colonization
- tend to be associated with disturbed ecosystems
Landscape effects of Establishment of invasive species
- Spatial configuration of habitat that promotes survival and reproduction (landscape effects on demography)
- tend to be associated with disturbed ecosystems
- Spatial distribution of disturbed areas, network, connectivity, to find vulnerable areas
Landscape effects of Dispersal of invasive species
- Spatial configuration of habitat that promotes movement of exotic species or their dispersal vectors across the landscape and thus affects the potential for invasive spread
- Propagules, long vs. short range, different mechanisms/abilities
- Invasives tend to be generalists and good at dispersing
Landscape effects of Spatially distributed populations of invasive species
Interaction of previous stages w/ landscape structure may give rise to spatially distributed populations that set the stage for invasive spread (nascent foci)
Landscape effects of Invasive Spread of invasive species
- Landscape ecological perspective may be required to predict and manage invasive spread if spatial distributions of habitat or resources affect any of the stages of the invasion process
Characteristics of invaders
- Generalists
- Tough competitors
- Allelopathy (broom)
- High reproduction rates (rats)
- Persistance of reproductive resources (seed bank)
Landscape effects of invaders
- on dispersal vectors by enhancing the spread above some critical threshold
- on promoting or altering species interactions in ways that enhance invisibility of communities
- Comprising of enhancing the adaptive potential of native species to resist invasion
- Interacting with disturbances in ways that cause resources to fluctuate, which can enhance invasibility
Why should ecologists care about the effects of landscape heterogeneity on plant species diversity?
- Disperse, establish, survive, reproduce
- Dispersal of plant portables across landscape could be affected by the spatial arrangement of patches
- Edge and dispersal, both facilitate and inhibit dispersal
- Many invasive plants like edges and small patches
Landscape effects on biodiversity of native species and non-native species (kumar et al., 2006)
- At landscape level, best models explained 43% of variation in native plant species richness, and 70% of variation in non-native plant species richness
- Variation explained was always higher for non-native richness, and inclusion of landscape metrics (heterogeneity, edge) always improved the models
Non-native species management at the landscape level
- Offense: Contain invaders at their source patch
- Defence: protect uninvaded destinations from invasion
Which is better to reduce overall spread rates of invasives? Offence or defence? When?
- Offense better for early invasions
- Defence better after increased invasion, and when goal is to protect high conservation value areas
What factors may affect choice of offence or defence for invasion?
- Spread mechanisms (do birds carry seeds long distances, streams are good at dispersing)
- Connectivity and disturbance by humans
Ecosystem services
- Nutrient cycling
- Pollination
- Habitat
