Overabundant species are considered drivers and passengers of profound anthropogenic disturbance in ecosystems, resulting in uneven communities. Understanding the causes of spread and establishment of such species may help decipher invasion mechanisms, while providing managers targeted management tools. The objective of this research was to quantify the spread of Corvus C. cornix in Israel, while attempting to elucidate the causes of its spread. Long-term occurrence data of crow sightings was used to analyze the species expansion. This data-set was correlated with a suit of climatic, NDVI and land-use variables. Results showed new hotspots of hooded crow activity in the north-western Negev desert and Jerusalem regions. A diffusion equation model revealed an expansion rate of 1.60 km year ^-1. Land-use analysis revealed an affinity of sighted individuals towards urban and low vegetation land types. Hooded crows sightings were positively correlated with annual precipitation, while being negatively correlated with precipitation during the wettest quarter. These findings suggest the species has established new source populations and is situated in the range expansion stage. The comparatively slow rate of dispersal is consistent with a neighborhood diffusion pattern, corresponding to the species life history traits. Human- altered land-use types, including low cover agriculture provide a habitat rich in constantly available food and nesting trees, both allow the hooded crow to thrive throughout the year. Precipitation may aid in enhancing hooded crow tolerance towards other unfavorable physical conditions. In light of these new findings, short term actions require pruning of tall trees in population hotspots. Intermediate term policies should focus on removal of existing stepping stones, and on farmer-ranger cooperation with the goal of limiting available food and water supplies. Long-term plans ought to recognize centers of hooded crow activity as indicators of highly disturbed and uneven communities. This should emphasize the need to establish Agri- environmental schemes (AES) in such areas, which would raise community resistance to invasive species. As AES are currently not in place in the national scale, and since their creation has the power to improve landscape connectivity of native species, this last plan is especially in need.

Count & Environmental Data

Crow sightings are periodically collected through bird counts done in the field by INPA staff, or randomly sighted and recorded regularly by INPA rangers. These sightings include GPS coordinates and are stored at the INPA data center. Our analyses included data from 3216 surveys out of 7517 sightings in total, and 4301 random observations, done predominantly since mid-2008.

Environmental data included NDVI data obtained from available governmental projects in the U.S. (United States Geological Survey, n.d) and a set of 19 climatic variables acquired from an open source initiative (Fick and Hijmans 2017).      

Data Analysis

Two time frames of ten years each were chosen, corresponding to a period of a sharp increase in crow counts, 1998- 2007 and 2008-2017. The data was mounted on a map using ArcGIS software (ESRI, 2018), and subjected to various types of analyses.

First, a point density analysis was conducted, in which the center of each cell in the map was assigned a 'neighborhood' of a certain size. Then, the sum of all crow counts in this area was divided by the neighborhood area size. The resulting value was represented on the output layer as a designated color.

In order to quantify the rate of spread, a reaction diffusion model was used based on the classical model by Skellam (1951) , which can be approximated as

R²=4rDt²                                                                

(Okubo and Levin 2001), where r is the population intrinsic rate of increase, D is the diffusion coefficient, and R is the radius of the area occupied by the species at time t. Since the square root of the area is proportional to t, a linear relationship exists between the two, and the slope of this linear line is

                                                      (2)

Demographic data of the hooded crow in Israel suggest that its intrinsic population growth rate equal to 0.09 year ^-1, i.e., an annual increase of ~9.4% in population size assuming an exponential growth. The resulting rate of spread was corrected by √2 . This was done to account for the ability of dispersing individuals to move only in the direction of half a circle, since Israel borders the Mediterranean Sea see: (Lensink 1997).   

The later time period was matched with a land-use layer with data obtained on 2016, and provided by the Israeli mapping service, and each crow count was assigned the land use on which it was observed.

 Another spatial tool which was used to infer on the causes of spread was the generalized structural equation model (SEM). The chosen time period for the model was selected to be 2013-2017, when the increase in crow sightings was highest. Hooded crow sightings of this period were extracted from the data. The NDVI and climatic variables were matched to each sighting, taken in one month intervals. This was done by dividing the country into a grid of squares. The size of the grid was chosen to be 3.5 square kilometers by running a partial least square regression including the climatic variables and NDVI scores, and selecting the grid which offered the highest R squared values. Values of minimum, variance and sum out of each sighting present between 2013-2017 (a time when the increase in crow counts was the highest) were calculated of each month, without summer months when there is expected to be no crow activity and climatic variables (6 chosen out of 19) were matched to crow sightings. Then, the strength of the effect of each explanatory variable and its directionality (positive/negative) on crow counts, and between climatic variables to NDVI, was assigned a numeric value.

Data uploaded includes crow counts done by the Israeli nature and parks authority (INPA) and a land-use GIS layer of the year 2016. This data is required in order to reproduce the research done in the paper.