Week 8 - Spatial Interpolation

Spatial Interpolation using three different methods.
Los Angeles County Rainfall calculation

Introduction

This exercise uses rainfall data from 58 Los Angeles County station for the purpose of creating some interpolated maps of the region about trends in precipitations and its distribution by interpolation of existing data.
The datasets available for rainfall for both current and average normal seasons in Los Angeles County Area are limited to 58 points, which have been manually entered and geographically located through X-Y coordinates definition and taken from the Los Angeles Water Resources Department under "Near real-time precipitation map" (complete list given below at bottom page).
The following list of datasets includes three different methodologies of interpolation analysis, each with different characteristics and quality results, able to give a clear idea of trends and current situation of rainfall for this area. These methods are the IDW method, the Natural Neighbor method and the Kriging method.

Rainfall: Current Season

The above map shows the rainfall as in 25th February 2011 for the current season.
No matters the method used for calculations, it is possible to see how the Eastern
area of the County, corresponding to the Los Angeles Padres Forest, has bigger amounts
of rainfall. This could be related with the fact that many of the stations in that area are located
at higher elevations, where altitude may have influenced amounts of rainfall and then
the amount of vegetation grown in that National Forest. On the other hand, desertic and internal areas have as expected the lesser amounts of rain. There is an augment in coastal areas where micro climates influence rainfall.
The Inverse distance method has as a result a map showing several "islands" of different values
within neighbor limits of those stations having higher values than the rest of the areas more conformal and balanced to the general trends. This can cause several cases of exceptions and gives more weight to the singular records, and so resulting in a map showing several exceptions and a general trend less conform.
The Natural Neighbor method is represented in map two. Here, the trends and the values look shared more equally within all the map, causing less prominence of singular differences compared with the previous map. I think it that in this case, given the fact that the distribution of rainfall within the territory has a simple spatial organization, this is the best of the three methods described for overall three analysis. The fact that one point is closer to a particular station has a strong relationship also in how close is the amount of rainfall to that same station.
Other factors calculated in weighed methods are less precise than this.
Kriging method shows in this case a precision that could be compared with the Natural Neighbor method, and creates Contour lines that are equally shared within the middle-points areas and finally very realistic.

Rainfall: Historic mean of Los Angeles County

The second map shows the rainfall of a normal season in Los Angeles County. It is inmediately clear that the third case, the Kriging analysis show a map of precipitation substantially different form the other two. This is because in the normal values have been created with a mean calculation process that didn't allow the record of higher peaks of rain in one single station (which have been recorded in more than one case between the current season records).
Kriging analysis had furthermore smoothed and flattened these values, creating a map overall more flattened and with no peaks of value.
By comparing the second maps with the first ones, it is possible to see how rainfall during this current season has been lower than the average historic rainfall. It is possible to recognize also in this case regions where vegetation is more likely to grow and desert areas at North where rainfall is extremely low.
Also in this case, Natural Neighbor Interpolation method produced more equally shared distribution and contour lines of levels of rainfall rather than the Inverse Distance Weighted.

Rainfall: Mapping the Difference within current and historical rainfall

The last map above shows the difference between the current amount of rainfall and the average records for the Los Angeles County area. Blue colors indicate higher rainfall than normal, while Red/Orange values are areas receiving less rain than usual.
Natural Neighbor and IDW methods are methods showing a similar pattern, both indicating as there has been an increase in Western and Southern areas of the County, which correspond to the Los Angeles Padres Forest and the Coastal shoreline between Venice and Long Beach.
Not much has changed instead for the desert areas North of the County. Most of the County however had received much less amounts of rain than expected and probably is affecting reservoirs closer to most urbanized areas. These are in fact located at South-West, where most of the red areas created by the interpolation are located.
In this case, the flattening resulted with the Kraging interpolation method had shared the rainfall more equally within the southern areas and has given probably the most defined map of the differences between the two datasets as a result of the Interpolation.

Table of Records

NAME	ID Number	LAT		LONG	Elevation feet	Current S.	Normal S.	Difference S.	Elevation mts.
Acton Camp Precip	58	34.450556		-118.198333	2,625.00	6.65	10.03	-3.38	800.10
Agoura Precip	56	34.135556		-118.751944	800	13.7	17.84	-4.14	243.84
Aliso Canyon Precip	57	34.328333		-118.554722	2,367.00	20.16	22.82	-2.66	721.46
Aliso Canyon Wagon Wheel	4	34.274722		-118.526389	1112	17.13	19.57	-2.44	338.94
Avek Precip	54	34.539167		-117.923056	2,825.00	6.06	5.78	0.28	861.06
Ballona Crk Precip	55	33.998611		-118.401389	38	8.94	12.39	-3.45	11.58
Bell Canyon (Rocketdyne) Precip	52	34.178056		-118.590833	2,260.00	14.02	14.74	-0.72	688.85
Bell Cyn Debris Basin Precip	53	34.194167		-118.656389	895	12.28	15.57	-3.29	272.80
Big Pines Recreation Park Pcp	50	34.378889		-117.688889	6,860.00	26.73	24.5	2.23	2,090.93
Big Rock Mesa Precip	51	34.039167		-118.616667	300	19.65	16.58	3.07	91.44
Big Tujunga-camp15 Precip	46	34.289167		-118.288056	1,525.00	13.74	19.36	-5.62	464.82
Bouquet Cyn @ Urban Precip	47	34.448333		-118.505833	1,300.00	9.88	7.98	1.9	396.24
Brand Park Precip	48	34.188056		-118.271944	1,250.00	14.17	18.34	-4.17	381.00
Brown's Canyon Precip	49	34.311667		-118.607222	2,400.00	19.25	19.55	-0.3	731.52
Cedar Springs Precip	42	34.355833		-117.873333	6,780.00	30.16	29.95	0.21	2,066.54
Chilao-St Hwy Precip	43	34.317778		-118.008056	5,275.00	20.75	22.9	-2.15	1,607.82
Clear Crk School Precip	44	34.276944		-118.17	3,150.00	27.52	30.54	-3.02	960.12
Cogswell Dam Precip	45	34.243333		-117.963333	2,300.00	25.47	34.21	-8.74	701.04
Domin Wat Co Precip	39	33.831389		-118.224722	30	13.9	12.11	1.79	9.14
DP WHQ	1	34.081667		-118.150278	466	18.5	16.82	1.68	142.04
Eagle Rock Rsvr Precip	40	34.145556		-118.19	1,085.00	16.34	18.14	-1.8	330.71
Eaton Wash Precip	41	34.074722		-118.054722	261	16.26	12.4	3.86	79.55
Flintridge Precip	38	34.181667		-118.185556	1,600.00	23.03	22.09	0.94	487.68
LA 96th & Cen. Precip	36	33.948889		-118.254444	121	12.99	13.78	-0.79	36.88
LA Ducommun St Precip	37	34.0525		-118.236667	306	13.54	15.63	-2.09	93.27
La Mirada Precip	28	33.883056		-118.016667	75	9.33	13.45	-4.12	22.86
La Rvr @ Wardlow Precip	29	33.819722		-118.122	25	11.3	9.9	1.4	7.62
La Tuna DB Precip	30	34.236667		-118.326667	1,160.00	12.52	16.44	-3.92	353.57
Lancaster Roper Precip	31	34.679722		-118.010278	2,400.00	7.24	5.34	1.9	731.52
Lancaster Waterworks	3	34.801111		-118.559722	2910	7.24	5.34	1.9	886.97
Le Habra Heights	11	33.960833		-117.950556	755	15.98	15.85	0.13	230.12
Lechuza Pat Sta Precip	32	34.076944		-118.879444	1,620.00	21.85	22.4	-0.55	493.78
Lewis Ranch Precip	33	34.42		-117.886389	4,615.00	17.17	15.21	1.96	1,406.65
Little Gleason Precip	34	34.378333		-118.149167	5,600.00	15.67	23.87	-8.2	1,706.88
Little Rock Crk Above Dam Percip	35	34.478056		-118.023333	3,267.00	9.8	9.16	0.64	995.78
Loomis Ranch Precip	27	34.348333		-118.048056	4,325.00	15.39	18.55	-3.16	1,318.26
Mescal Smith Precip	24	34.4675		-117.711111	3,810.00	9.29	7.8	1.49	1,161.29
Mint Cyn @ Fitch Precip	25	34.446667		-118.4275	1,652.00	9.02	9.05	-0.03	503.53
Monte Nido Fire Precip	26	34.077778		-118.692222	600	19.65	22	-2.35	182.88
Newhall-Sol Precip	21	34.425833		-118.578333	1,243.00	13.74	17.89	-4.15	378.87
North Lancaster Precip	22	34.761389		-118.125	2,310.00	2.68	5.19	-2.51	704.09
Northridge- LADPW Precip	23	34.231111		-118.541111	1,117.00	12.8	15.3	-2.5	340.46
Pacoima Dam Precip	20	34.33		-118.399444	1,950.00	17.2	19.54	-2.34	594.36
Palmdale Water Dist Precip	17	34.594722		-118.091944	2,595.00	7.17	6.93	0.24	790.96
Pine Canyon Patrol Station Pcp	18	34.673333		-118.429167	3,286.00	17.24	19.14	-1.9	1,001.57
Pnt Vicen Ligh Precip	19	33.741667		-118.410556	125	18.87	11.07	7.8	38.10
Pudd Div Precip	15	34.129167		-117.780833	1,130.00	20.43	19.21	1.22	344.42
Quartz Hill Precip	16	34.648056		-118.24	2,395.00	9.96	7.97	1.99	730.00
Redman Precip	14	34.764444		-117.925	2,360.00	5.59	5.18	0.41	719.33
Relay	2	34.735		-117.776944	3057	6.52	5.15	1.37	931.77
Rocky Buttes Precip	13	34.65		-117.863333	2,540.00	6.57	4.84	1.73	774.19
San Gab Pow House Precip	12	34.155556		-117.907778	744	16.16	22.98	-6.82	226.77
San Gabriel Dam Precip	7	34.205556		-117.860556	1,481.00	28.78	28.86	-0.08	451.41
San Gabriel East Fork Percip	8	34.235833		-117.805	1,600.00	25.55	26.04	-0.49	487.68
Sanberg Airways Station Precip	9	34.746389		-118.724167	3,635.00	10.94	12.85	-1.91	1,107.95
Santa Anita Dam Precip	10	34.184167		-118.019722	1,400.00	27.17	26.21	0.96	426.72
Tanbark Precip	6	34.205278		-117.760833	2,750.00	43.66	28.04	15.62	838.20
Topanga Canyon Patrol Station Pcp	5	34.084167		-118.599167	745	18.89	24.42	-5.53	227.08

Fire Hazard Risk Map - Week 7 - Spatial Analysis II

Mapping Fire Risk

FINAL MAP: FIRE HAZARD RISK AND STATION FIRE EXTENT.

FIRE HAZARD RISK MAP WITH WETLANDS AND ROADS

This Exercise show how to make a Fire Hazard Risk Map for the Area of Los Angeles County, considering the availability of different GIS Datasets and the powerful tools of ArcGis Spatial Analysis Extensions.

First, It has been collected three DEMs from the Seamless Server Dataset covering 34N117W; 34N118W and 33N118W. Have been unified into one single Raster Elevation Dataset as ESRI GRID.

The second step has been the calculation of relative slopes across the whole area.
The slopes of elevation have been calculated with the percentage mode, and divided into 9 significative ranges that represent strong differences in fires and gases behaves:

3%-6%-12%-20%-50%-85%-150%-300%

The average augment between one range and the following has been divided into 66% (four cases) and 100% augment (four cases) from the previous range.
This in order to distribute an exponential level of change within every class.

These ranges have then been reclassified considering the exponential range values into ten new classes, which combined for the from level 1 to level 8. This scale identify the advantage gained by fire in these areas, considering that steepness is a factor able to increase fires' power and speed, and that warmer gases tend to move in upper elevations of the atmosphere.
This values will weigh up to 8 points or 22.2% into the final hazard scale for the Station Fire Area, which has a total of 36.

The zone where Station Fire sprawled two years ago is an area of high elevation and steepness, which present no rivers or water land, differently form other areas in LA county.
(see map 1)

A second factor for the identification of Fire hazard risk factor has been the land coverage.
Using the scale produced by researchers from Greece (see Ref. 1), who determined risk indices for Mediterranean forest species based on flammability properties of Vegetation, and the Land Cover data from USGS Earth Explorer of Los Angeles County, the different coverages have been grouped into a classification scale with factors ranging from 0 to 25,
which will be added later to the original 1-8 scale for Slopes using Raster Calculation.

Also the presence of water in nearby areas has been taken into account for two reasons:
- First, wetlands act as natural obstacles to fire, and delimitate the sprawl with benefice consequences for the areas close to them.
- Second, in case of Firemen intervent, the availability of water ready to use to fight against fire in crucial areas not always reachable by airplanes and helicopters or when those are used in other zones, is an important factor.
Areas of Station Fire zone have been divided into four ranges corresponding to buffers of 5, 10, 15 and 20 kilometers from an available resource of natural water, from 1 to 4.

The fourth and last factor is the distance from available roads. Roads may act as cut fires,
can facilitate detailed intervents into particular areas, which otherwise necessitate the action of Canadair planes and helicopters. To this factor have been given values of risk augment from 0 (within 200 meters from an existing road) to 3 (more than 2 kilometers from an existing road).

Raster data Calculator allows analysts to weigh every factor easily, according to what other specific experts say regarding particular factors, such as the fuel potential of vegetation or the behave of fire with steepness factor.
The final map is the final result of these four different factors into a weighed, comprehensive map of risk.
The Station Fire occurred into an area matching with these factors, as is it possible to see in the last map. Many Red Squares are within its boundaries. Probably also factors of climate and predominant winds had an active role in determining the shape of the involved area. Cells with different color legend are missing one or more of the factors of calculation, for example are outside the buffer calculated for roads of wetlands, or still missing Land Coverage data or Slope of Elevation.

It is included again the road and wetland layers in order to show a geographic reference.

The power of GIS Analysis is extremely impressing.

MAP 1: Rivers and Wetlands within Station Fire Extent.

MAP 2: Creating a Fire Risk Grid
MAP 3:

1) Determining hazard risk indices for Mediterranean forest species based on particle flammability properties
S. Liodakis, a, , I.P. Agiovlasitisa, T. Kakardakisa, N. Tzamtzisa, D. Vorisisa and E. Loisa
a Department of Chemical Engineering, National Technical University of Athens (NTUA), 9 Iroon Polytechniou Street, Athens 157 80, Greece
Received 9 February 2010; revised 20 June 2010; accepted 23 November 2010. Available online 24 December 2010.

Suitability Analysis - Week 6

DISCUSSION

GIS, especially when combined with Raster data and remote sensing techniques, are amazing tools for identifying areas and locations most suitable for a given landuse or a particular activity.
As shown in the above maps, spatial analysis of complex problems such as finding the best place for several activities has never been so facilitated by technology.
A GIS can perfom several kind of analysis in a separate and/or combined manner. It is possible, for example, to quantify and apply particular buffering zones to the most important environmental resources such as forests, biospheres in danger, or simply protect wetlands from pollution. The wide range of personalized tools it is possible to apply, in fact, are able to satisfy almost an infinite number of spatial requirements such as the preferred buffer zones from a particular animal habitat or river. Thanks to the different formats of datasets: Vector, Raster and database, it is possible to visualize, measure, store and work on any typology of information needed by the analysis process.

In the case above, several variables have been considered in order to find the suitable locations for the construction of a new landfill in the territory of a county:

Calculed Elevations of Slopes, Terrain surface coverage, distance from existing landfills and wetlands, as well as terrain's permeability have been all studied and considered properly, then finally weighted according their real importance as a factor affecting the project.
This makes GIS an irreplaceable tool for Urban and Regional Planning, and should be always considered when performing complex spatial analysis, also because it may reveal patterns or relationships hidden at the bare human eye.

As an example of its powerful analysis capability, it has been recently put focus on the case of the expansion of a landfill in California's Central Valley: some people died and others had serious health trouble and birth defects for the contamination of the Drinking water system of Kettleman, a city in the Central valley which has high levels of agricultural and farming activities. the Element responsible for the deaths apparently is the Arsenic, highly used in pesticides, herbicides and metal alloys. A Landfill is currently located 3.2 miles south of the city, and until present times it has been used as waste disposal area for Mothball and other dangerous elements. It is likely that the Waste Management company built that facility in times when only surface features such as elevation were taken into account when deciding the suitable location.

There are many factors involved into the spread of polluting substances into environment and their interrelation with human activities. The reasons of this fatality may have been related with a wrong spatial analysis, which took into account exclusively the presence of people in the immediate nearby of the waste facility. In reality, the factors and the risks are several and different, and also have different weight in terms of importance. The presence of wetlands of subterranean rivers and the porosity of the soil are factors that can extend enormously the area of risk, and in a way which is not uniform but has influences from the slopes and the elevation of the area, or even the amount of rain within the site.
Only a well planned GIS analysis can take into account all these factors and establish a proper mathematical formula and model that takes into account all the known variables.

In the particular case of Kettleman City, a previous GIS analysis should have been conducted in order to establish if the location was suitable, considering the river systems, the permeability of the soils, and the typology of material disposed in that facility.
The Department of Toxic Substances, having the power to establish State or Nation level Guidelines and provide experts for GIS analysis, should establish a process of deep investigation to be addressed before every new waste management plant or landfill is created.
Using all the possible and known factors, including but not limited to:

- Climate and Rainfall factors;
- Soil permeability/drainage
- Distance from Wetlands and their eventually connected river system network;
- Land typology Coverage with qualitative differences between the surfaces involved;
- Elevation;
- Slopes of elevation;
- Distance from human settlements and activities including land used for farming, agriculture, ect.
would be possible to establish a standard GIS model and procedure that can be applied to all the locations nationwide or Statewide when deciding the suitable location for a waste treatment zone or landfill area.

The second most important advantage is that with GIS, using the Raster Calculator and the Reclassification tools, it is possible to change the importance of each singular factor to a scale that fits well the reality and the game of factors involved.
Examples of bad environmental management such as Kettleman City could be in this way avoided in future and the whole general procedures generally enhanced for a better and qualitative service to citizens.

Giuseppe Riccardo Belvisi

Buffering Marijuana dispensers - Exam

Dear LA City Council, I would like to express my agreement with the idea about the institution of the 1,000 feet buffer law from the marijuana dispensers.

I thought the idea was very unlikely to be realized because of the limites space in the urban tissue, but considering four kinds of locations:

- Schools;

- Churches;

- Recreation areas;

- Public Libraries;

- Public Parks;

Only 3 of these public facilities and institutions resulted within the 500 feet boundaries,

While more than 22 of these resulted in the requested 1,000 feet buffer, as you can see in the selection.

The reason is that I think as a Geographer that such an objective is practically impossible to realize. The buffers of all the locations involved, such as public libraries, schools, recreation centers.

I would rather suggest to transform the buffer into a 500 feet buffer zone from all those five locations. In this way the changes for retailers should be acceptable, rather than changing locations for all the 5 typologies of locations and the retailers/dispensers themselves

Reasons to modify the current law proposal of 1,000 feet buffering from Medical Marijuana Dispensers