Science topic

# Spatial Ecology - Science topic

Explore the latest questions and answers in Spatial Ecology, and find Spatial Ecology experts.

Questions related to Spatial Ecology

I explored species association using survey plot data in Tibetan Plateau, the highest plateau in world, where very unique herbaceous communities occurred. However, I found the number of associated pairs of species fluctuated based on a four-year interval data. Could I assume that the slow-growth community remain stable or equilibrium so that their association keeps stable too?

I have some environmental covariates derived from the digital elevation model (slope, gradient, channel network distance, etc.) in raster format.

**I want to identify areas of similarity between the covariates and somehow identify the smallest possible size area (or areas) to serve as a reference area.**

- Soil data points will be collected to create predictive models in this reference area.
- The predictive models developed in the reference areas should fit when extrapolated to regions outside the reference area.
- Therefore, the covariates will cover this external area.

I am looking for a solution with the R programming environment that will allow me to simulate animal movement (using a correlated random walk or other chosen model) within a polygon boundary, which acts as a reflective boundary to the movement.

I did find a solution (http://tinyurl.com/jbyuty8), but this has ArcGIS has a program dependency. I prefer to use open-source solutions.

The "adehabitatLT" package has a number of simulation functions, but I cannot find one that allows specification of a bounday argument.

Any helpful hints out there?

Hi,

I am checking for spatial autocorrelation in my dataset. It comprises the ID of the nests, the longitude and latitude for each of the nest boxes and the number of fledged chicks for each nest box. I want to know if reproductive success is spatially autocorrelated in our bird colony.

For this, I computed the distance matrix for nest boxes to know the distance between each nest box and the rest of nest boxes. Following this, I designed distance bands (distance lags) to calculate Moran's I for each lag specifically. As I have multiple data for several years (2014-2020), I wonder if there is any way to get a

*mean*Moran's Index of all the years, instead of calculating an index for each year.It is my first time doing these types of analysis so any advice would be very much appreciated!!

Thank you.

Iraida

Burlap traps are a way to mitigate the invasive

*Lymantria dispar dispar*(tussock/gypsy moth) caterpillars, which defoliate mainly hardwood deciduous trees. Burlap is wrapped around trees and tied with twine, then folded to create a flap and ideal conditions where the caterpillars migrate into. The caterpillars are then disposed of in soapy water when the traps are checked.If I want to study spatial ecology of these caterpillars, using quantitative analysis from each trap at a small lake surrounded by forest, how should I prioritize trap set-up (location, amount)?

Should the traps be completely randomized?

My study area is at maximum 2 square kilometres with a small Lake taking up about 0.25 of those square km.

Ideally I want to minimize confounding variables such as tree species the traps are placed on.

The goal of this project is to determine spatial distribution of the caterpillars and to mitigate them with weekly checks.

Any help would be greatly appreciated!

Hello,

I am currently trying to determine which interpolation method would best fit my GPS tracks. My "model" track has a 100 high quality consecutive locations with no gap and I excluded first random unique locations and then a number of consecutive points starting at a random position, to assess how accurate the different methods are.

I have tested linear interpolation (from move package) and correlated random walk using the method described in Technitis et al. 2015 (spacetime package).

I would like now to test a few curvilinear methods (e.g. exact cubic, natural splines, bezier...), but I can't find a spline or Bezier function that simply fills the gaps in my GPS tracks based on the missing timestamps. The spline function from base R collapses the results to a unique x value, which is not what I want obviously. Dr. Jed Long's package interpolatepathR would have been ideal if I only had a few unique fixes to interpolate, but I have far too many, and the package doesn't seem to handle long tracks (like my model track or longer).

So, does anyone knows of a package in R (I do not have access to Matlab) that handles spline and other curvilinear interpolation for (animal) paths, filling gaps (missing timestamps) from imperfect GPS data?

Thank you for your time,

Chloé

Hello, I have mapped the locations of several individuals of a certain species of bird. I have the GPS coordinates, the date and the time data every half hour for several consecutive days and I am trying to apply the Package 'mkde' to get the kernel density maps based on the movement (home range).
Is there someone with experience in using this package who can help me?

Thank you very much!

I was doing some geostatistical analysis (variogram+kriging) for a "presence only" type data in a species distribution modeling context. Since, we know that when estimating the (empirical) variogram, the attribute is basically assumed to be a realization of continuous random variables (although an attribute can occur in counts too). If the attribute is just the presence, and no sub-categories then all the values at all positions will be same (say 1, if we denote a presence by 1). Hence the variogram can not be calculated, not even the indicator variogram. In some papers such as

**[1]**and references there in, a grid based approach was used. In this approach a grid of certain size (e.g. 10 x 10 m etc) was superimposed on the sampling area and the number of species inside each cell were counted. This constitutes a count/frequency table like data. In the other approach**pseudo absences**or**background data**were generated using some algorithm e.g.**Maxent**etc (see e.g.**[2, 3]**). The pseudo absences are generated taking many factors into account and stacked/combined with actual data. This is merely generating x, y coordinates and giving it an absence status (say 0s). The result is a binary data with two categories, presence 1 and absences 0.Now the questions that are bothering me are

1. For the grid based approach, what should be the optimal cell size? How to find it and decide it? How to proceed with variogram with kriging etc?

2. For pseudo absences/background approach, how many absences (as compared to actual data)? How to decide it? How to proceed with variogram with kriging etc?

**Reference**

1. Rossi, Richard E., et al. “Geostatistical Tools for Modeling and Interpreting Ecological Spatial Dependence.” Ecological Monographs, vol. 62, no. 2, 1992, pp. 277–314. www.jstor.org/stable/2937096.

2. Tomislav Hengl, Henk Sierdsema, Andreja Radović, Arta Dilo, Spatial prediction of species’ distributions from occurrence-only records: combining point pattern analysis, ENFA and regression-kriging, Ecological Modelling, Volume 220, Issue 24, 24 December 2009, Pages 3499-3511.

Dear colleagues. Does anyone know of any work in which the criteria of functional traits of plants related to hydrological ecosystem services (e.g., hydrological regulation, drought mitigation or water saturation mitigation) has been used to spatially delineate an area for conservation planning, management, or decision making?

I think Moore et al. 2017 (Towards a trait-based ecology of wetland vegetation. Journal of Ecology 105: 1623-1635. doi:10.1111/1365-2745.12734) paper is enlightening from a conceptual aspect, but I need criteria and arguments from a spatial approach.

Thanks a lot

I need to calculate the Global and Local Moran indices in R using a variable distance as threshold: I have a SpatialPointsDataFrame with almost 300 points and I want to calculate the Global Moran index using 4 different distances (e.g 5 km - 10 km - 50 km - 100 km).

I read this discussion https://www.researchgate.net/post/Which_package_in_R_could_be_used_to_perform_Morans_I_test_for_spatial_autocorrelation

and I know the 'ape' and 'spdep' packages but it seems that no adjustment can be done concerning the spatial width to be considered...

Thanks

I would like to extract a number of bioclimatic variables (mainly from the Worldclim database) from the distribution ranges of several animal species. These distribution ranges are presented as shape files (.shp). Has anyone got some insightful links or info how to perform these analyses in an effective way using R? Unfortunately, I am only familiar with extracting variables from coordinates directly. Thanks!

I am looking for research and methodology for mapping roof-nesting seabirds in urban environments. This is for a pilot project. Any tips and articles are appreciated! Thanks.

Colleagues and I are weighing options on how to fit GPS transmitters to adult Black Swans in New Zealand. Adults weigh ~4-7 kg. We're apprehensive about using collars as they may get caught in vegetation while foraging. Also concerned about satellite uplink capabilities down here (collars are only available from overseas companies). Another option is a dorsal attachment, but we've been advised against using any type of harness. We're leaning towards tail-mounting ~40g transmitters (Sirtrack PinPoint Iridium) to tail feather(s) just after moult. We should be able to get ~3 fixes per day for 9 months, which would cover winter and the following breeding season. However, we're concerned about whether these will stay attached. Tail-mounting has been done on gannets, penguins, gulls... but I haven't seen this on swans or other large waterfowl. Curious if anyone has any suggestions.

Thanks in advance.

Mark

I am looking into Spatial Neighbors to address autocorrelation in my dataset, but I find it difficult to find arguments as to which method to prefer. I am using the R package "spdep" and functions dnearneigh and knearneigh to determine the distance-based neighbors and k-nearest neighbors, respectively. However, could someone advise me on the main differences between the two methods, as well as on how to determine d2 (upper distance bound) and k (number of nearest neighbors).

Exploring options (brand, make and model) for satellite transmitters to be attached to captured Northern Goshawks in northern Arizona.

Hi, I have mapped a data set where I have two populations - positive and negative for a certain bacterium (irrelevant to my question though). I used BatchGeo to simply map the data, no probem. Now I need to do cluster analysis of the points, or point pattern analysis, to see whether there is a pattern in the positive samples. I have tried to use ArcGISpro, and have mapped the points. I can't get any further though. I can't see how to label two distict groups (pos neg) and then analyse them. Anyone ever done this and can help? Or can anyone suggest a more user-friendly program to do so? I am sure ArcGIS can do this, but I have zero training in this and am stuck! I know nothing about coding and ArcGIS requires some knowledge, apparently.

Thanks for your help!

I want to compare the niche overlap in 4 species of arboreal lizards. The other issue I'm having trouble wrapping my head around is that in typical overlap analyses where habitat categories are used, they are discrete categories like "forest", "field", "wetland"... I have characteristics of trees that have sub categories such as bark roughness broken into 3 sub categories (low, medium, and high), tree complexity (low, high). This is because each tree has it's own habitat classification. In my data, different species prefer/avoid particular categories (based on manly selection ratios - adehabitatHS).

I'm wondering if I should do overlap analysis for each category (ex. bark roughness) separately because if I combine all categories (bark roughness, tree complexity, etc.), they % of use doesn't add up to 100% (although it does if I looked at tree complexity or bark roughness separately).

Hoping this is not too confusing, thanks for any input.

Dear all,

Does someone know if any R package can be used to perform meta-analysis take into account spatial and temporal autocorrelation (maybe separately)?

#spatial

I work on fish abundance data and associated diversity metrics at 35 stations located along several large French rivers (Rhône, Vienne, Loire, Meuse, Seine).

Some of the stations are closers than others : for example, 7 stations are located in the same area (distance <10 km) while some of them are located in different catchments without direct connectivity. Consequently, I expect that my data/results will be strongly spatially autocorrelated.

I am looking for a way to correct the time series meta-analysis for this spatial heterogeneity in R. Ideally, I was thinking of a method that would allow the weighting of the different time series in the meta-analysis according to their relative distance along the river network.

#temporal

The stations were sampled annually and the time series range from 18 to 36 years. So, consecutive years are likely to be more correlated than the first and the last years for instance. I would like to correct the temporal autocorrelation in the meta-analysis. For now, I have applied a Mann-Kendall trend analysis that account for the temporal autocorrelation, and I have extracted the correlation coefficient to be used in the meta-analysis. Do you think of another way to perform this correction?

1. I'm looking for GIS tools (software/plugins/etc) which are dedicated to analyze species habitat selection. Something that gives in the output habitat selection indexes or model/function.

2. Is there any way to estimate Resource Selection Functions in GIS software?

Thank You in advance for answers and comments!

Hello,

I would like to ask about any method for testing spatial dependence on categorical data (e.g. vegetation types polygons) and tools for modeling it against environmental data.

Is Multinomial Regression suitable for this task?

Thanks!

I'm trying to manipulate raster dataset using gdal software. When clipping bigger map to smaller raster data, it seems both gdal_translate and gdal_wrap work. What are the differences in using these two functions in general?

In my model, I would like to test how different plant characteristics (e.g. tree height, dbh, crown size) and their distance from the forest edge effects the fruit production. Several articles e.g.Bunyan et al. 2012 explain that the inclusion of multiple edges significantly improved their model fit based on AIC value. I would like to know, if I measure e.g.distances for the forest edge in the 4 cardinal directions for each tree, how could those four measurements be incorporated in one model, to test whether multiple edges have a higher effect than one-edge models.

Could you share references about works of remote sensing based on comparison between "Mean shift segmentation", "Watershed sgmentation" and "Multiresolution segmentation"?

I need to review the state of art in forestry aplications, vegetation delineation, trees detection, etc.

I would use the habitat model Invest software that requests as input data a series of weights for each land use on the basis of its suitability

Hello everyone!

What you think about the Evidence Likelihood transformation from categorical factors (geological, soils) to use the output as continuous predictors in Species Distribution Models?

Dear all,

I'm running the D and I overlap test metrics adapted for Warren, 2008 - 2010 in the ENMTools. According to Warren, D have ecological significance, while I test shows the comparison as a probability of distribution.

Have you used these metrics? how you explain the differences in the values of both tests?

Thank you very much in advance!

Ricardo

Am working on Spatial and temporal variability of rainfall in West coast of Karnataka using IMD gridded data of resolution 0.25 degree. Currently I did trend analysis for my study area and found drastic variation of rainfall in recent decades. I need experts suggestion to carry out further work for my research. Please help in this context.

Thank you,

I have 4 sites with a total of 22 species (i.e. site1 has 7 species, site 2 has 15, etc.). I also have multiple weeks’ species abundance data for each site. I want to analyze the species diversity on temporal and spatial scale based on high throughput sequencing.

Several methods have been proposed to compare sites for the species richness, many of which only use presence/absence data. I want to use abundance data, and do the following:

-use entire dataset, compare the similarity statistically, and obtain an optimum species richness/diversity value (say x number of species needed to reach 95% coverage of the whole dataset)

-use subsampled dataset (time-wise and site-wise), and analyze at which stage the previously obtained optimum number of species is reached (i.e., at 2 months of sampling instead of 12 or in one site instead of 4, etc.)

Any recommendations on the use of abundance data for answering these questions?

To avoid over-fitting, which variables should be removed first after preliminary MaxEnt run?

Is it those which are showing 0% contribution to the models, or those which are giving less than 0.8 AUC value when used in isolation (by examination of test jackknife data)?

Thanks

P

I'm looking for case studies of biological observations in floodplains that could follow the Intermediate Disturbance Hypothesis. The study noted below seems to be one of the only ones, but I'm guessing there should be similar observations looking at biota.

I am in a spatial ecology class and we are conducting marked point pattern analyses this week. If selfing rates aren't possible I would also be interested in fruit set or flowering.

Thanks

I am currently doing species distribution models using presence-only data, and I would like to perform spatial filtering of presence points in order to lower the sampling bias and to make sure that the data for building/evaluating models (I'm doing the data partition-based evaluation) are independent.

I have R function which leaves in one occurence point and removes other points in specified nearest-neihborhood distance area. However, choosing the distance within which the data will be rarefied is often arbitrary.

My concern now is how to justify or to select the distance threshold. As of now, I am using 1km threshold (duplicate points within 1km buffer are removed). My study area is ~250x50km, it is quite spatially heterogeneous, and resolution of my most coarse environmental variable is 1km (downsampled to 30 meters).

I feel that 1km threshold is somewhat adequate, but I cannot reasonabily justify this choice. Does anyone has tips on this issue or some articles to direct me towards. One method that I've came across in the literature so far is using variogram range where the points become spatially independent, but I am still not clear on how to use my environmental variables to build the semivariogram, so any tips here would be very appreciated.

I'm trying to estimate the degree of space sharing by pairs of individuals using the adehabitatHR package. After reading up on potential overlap metrics I settled on Bhattacharyya's affinity, but I'm finding that in some cases the home range overlap of an individual with itself, is greater than 1. I'd like to understand why this is the case, so that I can determine how best to deal with the resulting matrix! Thanks in advance for any help!

Hello,

I would like to ask the better choice for sampling a finite population on an area (both large enough, such as households in a political unit). I have seen some published papers and books (e.g. "Sampling Spatial Units for Agricultural Surveys") but I'm still aware of spatial dependence problem.

Any tip?

Thanks!

Hi everyone.

**I would like to know about the spatial analysis program(for Species Distribution Modelling).**

I tried using the MaxEnt and ArcGIS(ArcMap 10.1). MaxEnt is pretty useful. while using this program or software, I have question. I wonder if the new spatial analysis program Like MaxEnt.

If you know new(in my case... haha) program, please recommend!!!!!

ps1. i'm not good at R. if possible, please Recommend except for R package.

Thank!! You have to be happy.

How can I fix a leak from the pre-column side that is connected to injector?

what's the 'midpoint' latitude really meaning?

it is not very clear how to calcualte the midpoint of a species' latitude and longgitude when I try to pull together geography information from GBIF for species trait data. Generally, there are many distribution location records for one species, and only one latitude or longitude is needed in order to test effect of latitudial gradients on certain species' variable, i.e., plant height. sometimes I saw midpoint latitude was use in some paper but the methods was vague. centric or mean?

it seems there are risk to use midpionts in stead of exact site geographic locations, however, it is the only options in some circumstance, as sometimes the site location information is unavailable for many species in a compiled dataset.

any better way or comment? thanks.

Jingming Zheng

When adding the maximum background points(30.000) to the presence points (99.865) they don’t add up to the number of points used to determine the Maxent distribution (161.268 used when it should be 129.865). Or is it that the extra points are for about 20.468 different background points in each of my 3 crossvalidations? In this case why such a low number when my file as a few millions background points to sample from? Does Maxent decide that even if you tell it to use a max of 30.000 background points it doesn’t need that many?

The follow settings were used during the run:

66576 presence records used for training, 33289 for testing; 161268 points used to determine the Maxent distribution (background points and presence points); maximumbackground: 30000; replicates: 3

I am struggling conceptually with how I can best model my dataset. My data is the abundance of spiders collected from 20 permanent traps in a forested area. These traps have been repeatedly re-sampled every summer for ~15 years.

I have two sets of explanatory variables that I wish to model. The first is climatic data, namely rainfall and temperature. These have been obtained for each sampling year, so they are the same for all permanent traps. The second is geophysical data, including slope, elevation, aspect etc. These have been obtained through an online terrain model, so are unique to each permanent trap, but are obviously the same for each sampling year.

I'm trying to understand conceptually how to take these different scales of explanatory variables into account. Climate is regional, top-down whereas geophysical is local and bottom-up. I am after a relatively simple procedure if possible.

Any suggestions would be greatly appreciated.

Hi all,

My issue is that I am getting AUC values higher for a one explanatory variable model than for a multi explanatory variable model (that include the one from previous model). As my supervisor and I believe Maxent use all the info from the best variable and so adding variable could only improve the model, I have trouble to explain and this postpones a bit my thesis defense.

Could you give me an explanation here on how Maxent could allow such thing,

Thanks

It is suggested for reducing over-fitting of models, the variables should not be auto-correlated.

When each variable was tested through SAM module, it gave individual values of Moran's I which ranged from 0.2-0.6.

The question is how to know that that the variable has spatial auto-correlation?

Does Moran's I value of 0.2 with P level of 0.001 indicate auto-correlation?

and what is the cutoff limit for the value of Moran's I to select less or non auto-correlated variables for model development?

I’m looking for an index/metric that will permit to quantify the contrasting spatial distribution of a cover A (which is located mostly in the south of my study area) vs. a cover B (which is well distributed over the study area). I know that Fragstats can do that but I dont have the spatial analysis extension in my computer. Thank you for your answers !!

Hello to all,

I am doing a Phd in Population ecology using the Lesser Kestrels as a model species. I am using lighweight (~4 gr) GPS units from Technosmart to record the habitat selection and the movements during the Egg laying, Incubation and the Chick rearing period. I want to use a lighweight gps/gsm lightweight unit to track the juveniles leaving the natal areas and to monior recrruitment in colonies, unfortunately setting a base stationand radiotrack (due to time consuming) are either unappropriate methods . I am planning to track about 20 individuals and the area is bigger than 8.000 square kilometers. Does anyone has any idea?

Thanks in advance,

Konstantinos

What is the basis for a single species to occur in physiographically distinct landscapes (like montane and coastal).

Spatial prioritization for species recovery programmes needs to be evidence based, and it may be better to target effort in certain parts of a species rangeB but I'm struggling to find examples where interventions have been applied with good spatial replication across a species range. Agri-environment schemes perhaps provide the best opportunity for such studies. Maybe there are other interventions applied across big areas, that have been monitored and assessed (both the level of intervention and the animal/plant population response).

I would like to analyze data from different LTER sites in order to evaluate the effect of climate change on different ecosystems. The problem is that each site collected data with different study design and with different response variables (diversity of different taxa, variations in snow cover, biogeochemical cycles, ecc.).

I am working on spatial modelling of Himalayan terrain in connection to Climate Change and am required to make a spatial model and a decision support system. So what projection system I should use to get the best results am looking for habitat changes in fauna and flora as a whole.

I have created polygon shapefiles of vegetation patches (attached) and I would like to analyze them on the degree of clustering, orientation, directionality, etc. in comparison to a null model of randomly distributed patches. I am having difficulty finding software or scripts that will work for such an analysis (most specifically focus on point pattern analyses). I have tried 'spatstat' in R, but I am having no luck. Does anyone know of any tools that could help me? I would prefer any that may be free/open source, or can be used in common statistical programs (R, JMP, SPSS, etc.)... the more accessible the better!

Thanks!

When you run a simple mantel test, do the distance matrices have to be in the same order and to correspond the rows to the same individual or register?

I am working on a study of the spatial distribution of leprosy in an armadillo population. We have point locations for each captured individual and a binary measurement of whether each individual did or did not test positive for leprosy. This sort of marked point pattern is a form of univariate labeled data and I want to do a network (cross) Ripley's K analysis with a conditional randomization test, where the point locations are held constant and the mark (leprosy: yes or no) is permuted among the locations (rather than permuting the locations themselves). I have tried both SANET and GeoDaNet, which do a network version of a cross K analysis, but the randomization scheme seems to be for true bivariate data (not univariate labeled data).

Does anyone know of software that can do a "conditional randomization" test for network (cross) Ripley's K? Our software (PASSaGE 2) can do 1D, 2D, or 3D Ripley's K with this sort of permutation scheme, but not network K.

I need suggestions / help to the identify suitable model for spatial patterns and trends of forest succession. I have 5 land cover layers with an increasing trend of forest. Which Spatial – Temporal model is suitable? I want to use at least 4 layers for model input. I am confused by which one is better, GEOMOD, CA-Markov, ANN or another. Please help.

Does anyone have a good source for correcting for spatial auto correlation when comparing a species assemblage (site X species matrix) to a geographic distance matrix? I know a mantel test will tell me how correlated the variables are but how do you correct for this effect in subsequent analyses?

Given only animal relocation data from a telemetry study and no corresponding mark release recapture records is it possible to derive a density estimate using a mark release recapture model. I suspect there is a way to do this treating telemetry encounters as resighting events but I am not certain so I thought I would ask. If some of you can point me toward the relevant papers it would be greatly appreciated.

PAZ

I'm working with MAXENT and GARP in species distribution modeling (plant species). I've converted my raster files (environmental variables) to ascii ( I work with Arcmap 9.3 for converting) and then used them in MAXENT but now I don't know how convert them for GARP.

Thanks

I have mean-variance plots of annual NDVI from different data sets. I would like to compare the trajectories of these plots, but not sure what would be the best way to do it. Any suggestions are welcome.

Currents and other hydrodynamics / hydrological phenomenon can induce spatial expansion of a bloom (such as visible in upwelling areas).

I would like to know if abundance of cells can also induce a spatial expansion? As though phytoplankton needed "more space" when cells are more numerous?

The model I'm referring to utilizes the Markovian dependence of animal sign detection. Any opinions? Any pros or cons? Perhaps an alternative model?

Any tips on the best way to get sub-meter GPS coordinates while on the ground in Bolivia?

We are planning a study on roe deer migrations and habitat utilization and we think that GPS tracking of animals could be a part of this study. Since we have no experience in this field, could you please give recommendations on the set of equipment necessary to organize this work, and maybe any comments on the companies that sell such equipment?

Or maybe you know somebody who wants to sell used collars? Or somebody who owns the collars and would like to do research in western Siberia?

Any comments are appreciated! Thanks advance!

A coworker is working on a biodiversity habitat model with a minimum grid size of 30 x 30m. Basically, I am curious to know what the minimum habitat size a coot needs is and if the habitat is suitable to successfully reproduce since I am unable locate a definite number within the literature.

I'm doing research using shape metrics of islands and I'm trying to measure their length. However, some islands have rather odd shapes and I'm not sure what their length mights be.

Here is a link, just for example, on a map with three of the researched islands

I have a spreadsheet with the number of presence records of my study specie, sampled in different quadrants, of which I have the geographic coordinates of latitude and longitude. Whats the most appropriate way to analyze this data?

Does anyone know about courses of spatial ecology, landscape genetics, seascape ecology scheduled for this year? With application period still open.

I couldn't find any clear facts about the disambiguation between these two fields.

Hump shape distribution of biodiversity is the most common in the mountains. The mid-domain effect (only relatively simple model which can produce such a pattern) seems not as a driver of that pattern as it has no biological meaning. What about other hypothesis?

I need some articles related to this topic and some suggestions to analyze these data together.

Suppose we want to assess the effect of a given predictor variable on the given response variable. In case observations are dispersed in space it is recommended to check for spatial autocorrelation. But positive autocorrelation between points normally means that the values of the predictor are similar, and lack of correlation will be related to significant variation in predictor variables. Then, it is worth searching mainly for cases of negative spatial autocorrelation, when we fail to observe the effect of the predictor. Is this correct?

Can anybody give an example of empirical studies illustrating the effect of positive autocorrelation not related to variation in predictor variable?

I've done some looking on line, so far without success. For big or repetitive processing tasks that need spatial stats, it would be really useful...

The attached figure is a mark correlation function produced by the spatstat package in R. The summary returned from the object used to produce this figure is as follows:

r theo trans iso

Min. : 0.00 Min. :1 Min. :0.1663 Min. :0.1678

1st Qu.:15.27 1st Qu.:1 1st Qu.:0.1759 1st Qu.:0.1765

Median :30.53 Median :1 Median :0.5172 Median :0.4858

Mean :30.53 Mean :1 Mean :0.5695 Mean :0.5652

3rd Qu.:45.80 3rd Qu.:1 3rd Qu.:0.9435 3rd Qu.:0.9286

Max. :61.06 Max. :1 Max. :1.1325 Max. :1.2666

When I enter the name of the object on its own, I get the following:

lo k[mm][lo](r) lower pointwise envelope of k[mm](r) from simulations

hi k[mm][hi](r) upper pointwise envelope of k[mm](r) from simulations

I would like to return the actual values, as in the former output, for the simulations labelled "lo" and "hi" in the latter output.

In other words, I would like to be able to obtain a range of values at which the observed (black) line falls outside the grey envelope.

Is this possible?

I get the impression it may have something to do with the argument "savefuns", but I am at a loss to know how to implement it and return the values I need.

When analysing relationships between two spatially autocorrelated variables using linear regression (with ordinary least squares), one has to be aware of the inflated Type I error - you are more likely to get significant results even if you should not (e.g. Legendre 1993). On the other side, it seems that regression coefficients are not biased, so to estimate the slope of regression is safe. But how is it with explained variation in this regression, i.e. R2? Is it also inflated, similarly to Type I error rate? I read the paper of Lennon (2000) in Ecography, pointing out that R2 will increase with increasing spatial autocorrelation of one of the variables. But I can’t find other references dealing with this issue (there are plenty dealing with Type I error rate and shifts in regression or correlation coefficients, but I am not aware of other one dealing with R2). Can somebody please help me with that?

We want to select a fixed number of deer relocation, coming from collared animals, to sample the vegetation.

1- Do the sampled locations need to be spatially independent? Temporally independent? Spatio-temporally independent?

2- If so, how do you test for the independance. We were currently exploring autocorrelation indices (Moran's I) or spatio-temporal indices (Griffith's).

Among the methods available to select spatial filters (PCNM) which take spatial autocorrelation into account, which one is the most effective? Is there a consensus on this subject?

The general "nursery habitat" paradigm describes habitats that are distinct to juveniles (because they provide protection from predation, have ample food for lower trophic levels, and suitable environmental conditions). These habitats are used until individuals attain a certain size/age, and then movement to sub-adult/adult habitats occur.

It appears, however, that a single "nursery habitat" definition for a species is appropriate, as species-habitat relationships can change with ontogeny, multiple times even within the first year of life for an estuarine-utilizing fish.

Furey, N.B., Rooker, J.R., 2013. Spatial and temporal shifts in suitable habitat of juvenile southern flounder (Paralichthys lethostigma). Journal of Sea Research 76, 161–169.

What factors about a species life history or its environment mediate or influence how quickly species-habitat changes occur? Would we expect these changes to occur at specific stages, or more across a gradient (gradually)? And if species-habitat relationships are dynamic or even fluid when an organism is a juvenile, how does this impact how we manage habitats? Is this another case of focusing on landscape-based processes rather than habitat-scale? How does this impact the notion of Essential Fish Habitat?

e.g. when different scales for estimating plant cover have been used to assess plant species abundance in different experiments in the field, but then for reasons of wanting to compare across experiments, one would like to transform one scale to another (eg. modified Braun Blanquet to modified Londo)?

The different scales were used in different experiments since the main questions being asked required slightly different levels of precision, but now it would still be nice to be able to compare cover values across experiments.