I am not sure I understand your question, but, comparing CPUE between gears is tenuous.  The two gears likely have different size selectivity and fishing efficiency, among other things.  You can say one gear caught more than another, but if you use one gear in still water and the second gear in the rapids, you likely cannot treat the difference between the catches as representative of anything but a difference between the gears. 

Thanks for the response sir Cheree Steward.

I'm aware the differences between active and passive gear to the CPUE. 

So, the data that I have gathered (fish net and gill net) can't be presented together as I want to do GIS mapping based on CPUE and population distribution?

You might reduce noise from the data selecting some proper size thresholds (to take into account different selectivity). Not  sure if you would improve very much the results but what you could do is to work on GIS but not using absolute values. I think in this case you might  standardize them into the range 0-1 or coding them into categories (e.g. low- medium-high) using for instance percetiles. Than you could have two different maps (per gear/env. condition) that you could present together or, maybe better in parallel to check for consistency in the general patter. If you have samples/stations spatially and temporarily (overlapped) collected with both the gear you might also check for consistency in the signals and, perhaps (but in most cases very diffucult), check for an intercalibration process (i.e. assess the correspondence of threshols between categorial data between gear).

As I understand, you want to do is to map the abundance of fish in different habitats, to understand it's movements. So in my opinion you must use as many fishing gears as you can, as long as you use a standard unit to plot the data. If you know the effort made with each net, in your case I think you can use individuals/net area/time as abundance unit, i.e. Ind/m2/h, or ind/m2/min, depending on how long the nets have been used. With this unit you can plot or use your data together to know fish distribution.

We used a complex of fishing gears for estimating abundance of freshwater salmonids and graylings in streams and rivers of Russian Far East. We found that data collected by different tools is incomparable. We used casting nets, beach seine, gill nets (fixed and drifting), fishing nets and electro fishing. The usage of any tool was very specific and effectiveness depends of conditions a lot.

Gill nets has strong selectivity. So we should use a lot of mesh size of them for catch most of size classes. They are effective but it must be used in biotops without or very slow current or in part of river with clear bottom with gravel or sand in case of using as drifting nets in conditions of flowing water. In first case CPUE wich we used is number of individuals and biomass on square of net in time of exposition (ind (kg)/m2 of net/time unit). In the second case for CPUE estimation should be used area of river bottom were nets drifted (ind (kg)/m2 of bottom).

Beach seine can be used only with circumstances of clean bottom without stones and debris. Fishing nets and electrofishing is effective in case of catch at near shore wood debris and weeds and in relatively shallow places. Casting nets can be used in conditions of relatively clean bottom in both still and flowing water. CPUE  of all of this tools can be described as  ind (kg)/m2 of bottom.

But comparison of catch structure  of this tools is impossible. The catchability of them is too different and depends a lot on conditions of usage, species and size structure, seasonal and weather factors etс. So, for finding clear results of species abundance and distribution we tried to use only one tool which can be used in the most types of biotops.      

As previous answers comment, mixing two different sampling methods would lead to erroneus results due to mixing factors. As each net can be selective for a different size class or species behaviour, or have different efficiency, both metods probably will produce different results. You would not know if any difference found is due to a real difference between localities or to the sampling method. To evaluate the magnitude of the differences you should perform a previous study using both methods at the same time and the same places or use randomly each other in both sites during the study so that the variability due to the sampling method will be included as random factor. The first option do not permit to made statistical comparisons but you could at least to discuss the results. The second option will reduce the possibility of finding significant differences as the variance of the data increase, but would lead to correct results. You can use one or other method depending of the nature of the limitations to use one or other methods and if this limitation do not affect to all sampling localities but only some of them. I include here a reference to a paper that just compare different sampling methods for fish studies in coastal lagoons.

Assessment of fish assemblages in coastal lagoon habitats: Effect of sampling method

Por: Franco, A.; Perez-Ruzafa, A.; Drouineau, H.; et ál..

ESTUARINE COASTAL AND SHELF SCIENCE Volumen: 112 Número especial: SI Páginas: 115-125 Fecha de publicación: OCT 20 2012

The only way to avoid the evaluations described avobe by Angel is that you had been used both gears with the same effort in all habitats. I Thought that was the case,  so I suggested that you can use both data to map. 

Definitely not with different gear, probably not with the same gear.  Read Imperfect detection and the determination of environmental flows for fish: challenges, implications and solutions

Article in Freshwater Biology · September 2015 by Daniel C Gwinn et al.  I got my copy via this site: https://www.researchgate.net/publication/282314101_Imperfect_detection_and_the_determination_of_environmental_flows_for_fish_challenges_implications_and_solutions#full-text

I used mark recapture to estimate pop size of longfin eel.

Article Imperfect detection and the determination of environmental f...

Keep in mind that using a CPUE is always a way to percieve abundance when absolute numbers or biomass cannot be obtained. The CPUE is therefore an index that reflects abundance through a parameter of proportionality oftentimes called q. What you want then, is that your CPUE truly reflects the critter's real abundance modified only by the effect of q. You therefore want any other effect remove from your index so that a series of indices can effectively show how abundance changed in time or space. Removing these undesirable effects is achieved by standardizing your CPUE, most commonly by means of using Generalize Linear Models (GLM). By standardizing your CPUEs, it is very likely you can combine different types of indices and achieve what you are looking for. I recommend you use Google Scholar and look for the papers by Mark Maunder, use keywords "Maunder" "standardize" "cpue", they'll pop up right at the top. The paper, Methods for standardizing CPUE and how to select among them" by Hinton and Manuder (2003) will probably be most helpful.

I think the year for the Hinton and Maunder paper is 2004...

Standardizing catch and effort data: a review of recent approaches

MN Maunder, AE Punt - Fisheries Research, 2004

http://cse.fra.affrc.go.jp/kiyo/home/info_ex/projects/H18zero/MaunderPunt2004.pdf

Dear Radhi Amonodin,

How do you track the movement of fish? I mean, what types of tag and marker help you most in this regard whether its manual tags or molecular marker?

You can sample with two different gear types, but typically a CPUE is only calculated for each individual gear type.  There can be differences in selectivity, how the gear is deployed, etc.  If you want to compare the two side by side you'll have to standardize your nominal CPUE indexes and then compare them with trend analysis.

You can track movement, of course, but you can't compare CPUE between different gear, if selectivity is different, as it usually is

Which type are you fishing water. If you will surround fisheries area, may be you can try catch and release method, or zipping method etc...since the equal time.

Thanks Dan Gwinn, certainly the paper by Maunder and Punt is also quite useful and is probably the follow up of the initial report I was referring to (below). In any case, the message is the same, it may be possible to use CPUEs of different nature to compare trends in abundance or insert them in stock assessment models if properly standardized. Also, Mark Maunder has produced several papers that are quite useful to address the question under discussion here and many are freely available in the web.

http://www.spc.int/DigitalLibrary/Doc/FAME/Meetings/SCTB/16/MWG_7.pdf

Thanks a lot! I'm truly appreciate the answers and responses! I have understood that it's impossible to combine different capture tools since the units will differ accordingly. 

Nahid and Soner, I'm using removal method in a several sampling points as the species (Thynnichthys thynnoides) are known to migrate upper stream during spawning season. The capture recapture method are not favorable for this species since it's have been fished selectively during spawning season by the villagers and the catch up to 40 tonnes! Overfishing is one of the problems faced by the species. (refer to fish landing figure attached below). 

Thanks to Sasa for the recommendation to separate the maps according to gears/environmental condition.

No no... it isn't impossible! It is actually quite possible through standardization, look at the references above!

Thanks a lot for the reference Sir Carlos!